Download Distribution and Feeding Ecology of Bathylagus euryops (Teleostei: Microstomatidae) Along the Northern Mid-Atlantic Ridge From Iceland to the Azores

DISTRIBUTION AND FEEDING ECOLOGY OF BATHYLAGUS EURYOPS (TELEOSTEI:
MICROSTOMATIDAE) ALONG THE NORTHERN MID-ATLANTIC RIDGE, FROM
ICELAND TO THE AZORES
A Thesis
Presented to
The Faculty of the School of Marine Science
The College of William and Mary in Virginia
In Partial Fulfillment
Of the Requirements for the Degree of
Master of Science
By
Christopher J. Sweetman
2012
DEDICATION
I dedicate this work to the memory of my grandfathers, Charles F. Sweetman (1918 - 2011) and
Richard N. Langevin (1935 - 2008), whose strength and wisdom continue to be a source of
inspiration. You are both missed every day and I’ll do my best to try and “save the oceans”
during my lifetime.
iii
TABLE OF CONTENTS
Page
PREFACE ........................................................................................................................................v
ACKNOWLEDGEMENTS ........................................................................................................... vi
LIST OF TABLES….................................................................................................................... vii
LIST OF FIGURES…................................................................................................................. viii
THESIS ABSTRACT......................................................................................................................2
GENERAL INTRODUCTION…................................................................................................... 3
CHAPTER 1
Distribution of Bathylagus euryops (Teleostei: Microstomatidae) along the northern MidAtlantic Ridge.................. ..................................................................................................13
Abstract..............................................................................................................................14
Introduction........................................................................................................................15
Methods..............................................................................................................................17
Results................................................................................................................................23
Discussion..........................................................................................................................26
References..........................................................................................................................33
CHAPTER 2
Feeding ecology of Bathylagus euryops (Teleostei: Microstomatidae) along the northern
Mid-Atlantic Ridge............................................................................................................46
Abstract..............................................................................................................................47
Introduction........................................................................................................................48
Methods..............................................................................................................................52
Results................................................................................................................................58
Discussion..........................................................................................................................62
References..........................................................................................................................72
THESIS CONCLUSIONS.............................................................................................................87
VITA..............................................................................................................................................92
iv
PREFACE
Chapter 1 of this thesis will be submitted for publication in Deep-Sea Research I and Chapter 2
will be submitted for publication in Marine Ecology Progress Series, and each are formatted
under the guidelines of those specific journal.
v
ACKNOWLEDGEMENTS
This work was made possible by the gracious and dedicated efforts of many individuals.
Particularly, I extend sincere thanks to my advisor, Dr. Robert Latour, for his guidance and
support over the last year. I was not exactly a ‘traditional’ student and I appreciate your
willingness to take a chance on me and your extreme patience as I pulled this project together.
Additionally, I would like to thank Mike Vecchione for allowing me to join the FSV Henry B.
Bigelow expedition to the Mid-Atlantic Ridge. It truly was a once in a lifetime experience that I
will never forget. I am also grateful for the contributions and advice provided by each of my
committee members including Dr. Eric Hilton and Dr. Deborah Steinberg.
Collections of laboratory and field samples were made possible by the MAR-ECO
research crews, to whom I am eternally grateful. Particularly, I would like to thank Tracey
Sutton, Odd-Aksel Bergstad, Tone Falkenhaug, and John Galbraith, and other field personnel for
their support and knowledge of all things deep sea. I thank the crew of the R/V G.O. Sars and
the FSV Henry B. Bigelow for their excellent shiptime services. Post-cruise gut content work at
the Bergen Museum was supported by grants from the Sloan Foundation/Census of Marine Life
and I would like to thank Ingvar Byrkjedal for being a generous host during my visit. The data
analysis and synthesis phases of this study were supported by the NSF Ocean Sciences Division
– Biological Oceanography Program (OCE 0623551) and the VIMS Office of Academic Studies
and I thank our Associate Dean, Linda Schaffner, for always keeping an open door and for her
support during more chaotic times. Jeanna Hudson and Vanda Carmo were critical to the
identification of the various prey groups observed in the gut content analysis and their
knowledge proved invaluable throughout this project.
I would like to thank all of my friends that made these last few years an unforgettable
experience. This journey would have been far less entertaining without the help of all my friends
here at VIMS: Chip Cotton, Matt Whalen, Andrew Wozniak, Pat Lynch, Kathryn Sobocinski,
Mark Nelson, Kristy Hill, Cassie Bradley, Pat and Patty McGrath, Andrij Horodysky, Todd
Clardy, and Ali Deary. Also, I am tremendously grateful to my parents, Jim and Claire, and
brother and sister, Josh and Ashleigh, for their love and encouragement. I am also forever
thankful to my grandparents, some of whom are no longer with us, for pushing me to follow my
dreams. In particular, I thank my best friend and loving girlfriend, Kristene, for her support and
patience throughout this whole process. You assisted this project immeasurably, both directly
and indirectly, and I only hope that I can offer you the same support throughout your endeavors
in life. Without you, I would’ve never made it out alive.
“So long and thanks for all the fish!” – Douglas Adams
vi
LIST OF TABLES
Table
Page
CHAPTER 1: Distribution of Bathylagus euryops (Teleostei: Microstomatidae) along the
northern Mid-Atlantic Ridge
1. Fit statistics associated with the seven models fitted to Bathylagus euryops catch-per-uniteffort data............................................................................................................................38
2. Parameter estimates and calculations for the distribution model with the best fit containing
depth and ridge section as parameters.................................................................................39
3. Parameter estimates and calculations for the diel vertical migration model containing
depth and time of day as parameters...................................................................................40
4. Parameter estimates and calculation for the ontogenetic vertical migration model
containing dish size and depth as parameters......................................................................41
CHAPTER 2: Trophic ecology of Bathylagus euryops (Teleostei: Microstomatidae) along the
northern Mid-Atlantic Ridge
1. Zooplankton body types, corresponding geometric shapes and calculations used to
calculate the weight of prey in the gut contents of Bathylagus euryops.............................78
2. Gastric evacuation rates and daily ration estimates of Bathylagus euryops at different
temperatures along four different ridge sections of the Mid-Atlantic Ridge......................79
vii
LIST OF FIGURES
Figure
Page
CHAPTER 1: Distribution of Bathylagus euryops (Teleostei: Microstomatidae) along the
northern Mid-Atlantic Ridge
1. Sampling locations for Bathylagus euryops during the 2004 R/V G.O. Sars expedition
along the Mid-Atlantic Ridge..............................................................................................42
2. Sampling locations for Bathylagus euryops during the 2009 FSV Henry B. Bigelow
expedition along the Mid-Atlantic Ridge............................................................................43
3. Predicted catch of Bathylagus euryops at each depth zone relative to the four ridge
sections sampled..................................................................................................................44
4. Predicted standard length of Bathylagus euryops relative to depth zone at the CharlieGibbs Fracture Zone............................................................................................................45
CHAPTER 2: Trophic ecology of Bathylagus euryops (Teleostei: Microstomatidae) along the
northern Mid-Atlantic Ridge
1. Sampling locations for Bathylagus euryops during the 2004 R/V G.O. Sars expedition
along the Mid-Atlantic Ridge..............................................................................................80
2. Percent composition by weight and frequency of occurrence dietary indices for Bathylagus
euryops................................................................................................................................81
3. Size frequency of Bathylagus euryops used for dietary analysis........................................82
4. Scree plot and dendrogram from cluster analysis depicting dietary interrelatedness within
different size classes of Bathylagus euryops.......................................................................83
5. Canonical correspondence analysis biplot for Bathylagus euryops along the northern MidAtlantic Ridge......................................................................................................................84
6. Percent composition by weight of Bathylagus euryops A) smaller than 95 mm SL, B) 95 –
155 mm SL, and C) larger than 155 mm SL.......................................................................85
7. Percent composition by weight of Bathylagus euryops relative to the northern ridge
sections and the southern ridge sections..............................................................................86
viii
DISTRIBUTION AND FEEDING ECOLOGY OF BATHYLAGUS EURYOPS (TELEOSTEI:
MICROSTOMATIDAE) ALONG THE NORTHERN MID-ATLANTIC RIDGE, FROM
ICELAND TO THE AZORES
1
ABSTRACT
The northern Mid-Atlantic Ridge, from Iceland to the Azores (MAR), ranges in depth
from 800 – 4500 m and extends over an area of 3.7 million km2. Recent evidence from MARECO, a Census of Marine Life field project, reported increased abundance and biomass of deeppelagic fishes below 1000 m on the MAR. Among the fishes sampled, Bathylagus euryops was
found to be the biomass-dominant species and ranked third in total abundance. In this thesis, we
characterize the distribution and feeding ecology of B. euryops as a function of physical,
biological, and life history parameters along a mid-ocean ridge system. Multiple biologically
plausible general linear models were fitted to B. euryops catch-per-unit-effort (CPUE) data to
investigate the role of various combinations of explanatory variables on the distribution of this
species. Results indicated that a model containing categorical depth and geographic location
variables provided the most parsimonious description of B. euryops CPUE data. Vertical
migration analyses were also conducted to investigate the vertical distribution of B. euryops
along the MAR and results indicated that time of day had little influence, whereas ontogeny
likely influenced the vertical distribution of B. euryops. To describe the feeding ecology of B.
euryops, a general diet composition was determined. Multivariate analyses, including a cluster
analysis and a canonical correspondence analysis, were utilized to investigate factors that cause
variability within the diet of B. euryops. Results revealed that fish size and geographic location
significantly influenced the diet of B. euryops. Furthermore, daily rations were estimated to
better understand the role of B. euryops in the food webs of the North Atlantic and results were
consistently estimated to be less than 1% of the average wet weight along the MAR. The general
trend observed was a southward increase in daily ration estimates along the MAR.
2
GENERAL INTRODUCTION
3
The Atlantic Ocean is the second largest of the world’s oceans with an area of 87.0 ·106
km2, or roughly half the size of the Pacific Ocean (Jennings et al., 2001). Depths average 3600
m and are deepest in the Puerto Rico Trench (9560 m). The Atlantic is the only ocean with a
major connection to the Arctic Ocean, and hydrography is greatly affected by inflows of the
Mediterranean Sea, the Caribbean Sea and the Indian Ocean. The inflows to the Atlantic are
balanced by the formation and subsequent outflow of North Atlantic Deep Water, which supplies
deep water to all other major oceans via thermohaline circulation. The geological morphology of
the oceanic basin in the Atlantic Ocean plays a key role in determining its ecological
characteristics. For example, the Mid-Atlantic Ridge blocks the northward spread of Antarctic
bottom water and essentially modifies the hydrography of the whole Atlantic (Angel, 1997). The
scale of the Mid-Atlantic Ridge (MAR) is so large that, in order to describe, analyze, and predict
relationships among the organisms within it, a sensible approach of dividing the whole
ecosystem into ecological subsets must be taken (Herring, 2002).
The strongest and most predictable gradients in the open ocean are in the vertical
dimension. Many fundamental ecological processes are strongly influenced by the vertical
structure of the water column and the associated relationships with depth are typically complex.
Vertical gradients are usually related to the density structure of the upper water column. The
vertical variability of the open ocean has led to its division into three main oceanic zones,
defined here as: the epipelagic zone (generally surface to 200 m), the mesopelagic zone (200 to
1000 m), and the bathypelagic zone (1000 to near the bottom). Interfaces between zones tend to
fluctuate with time and space, are neither abrupt nor solid, but ultimately reflect biological
responses to varying environmental conditions. With the exception of chemoautotrophy in deep
hydrothermal vent or cold seep systems (Ramirez-Llodra et al., 2010), all primary production
4
occurs in the epipelagic zone, where seasonal variations are evident (Haedrich, 1997). Light
from the surface penetrates to about 1000 m, but there is too little for plant growth below 200 m.
The mesopelagic zone encompasses most of the primary thermocline, where conditions become
much more stable and invariant compared to the surface layers. Water temperature is relatively
constant and very low in the mesopelagic zone and nutrient concentrations are typically high.
The bathypelagic zone, however, is apparently too deep for sunlight to play a significant role in
the behavior and distributions of organisms (Angel, 1997).
Primary production varies greatly in the Atlantic Ocean but in general is extremely high,
around 13.7 Pg C per year (Behrenfeld and Falkowski, 1997), and is acutely seasonal. In winter,
surface waters are cool through a lack of solar heating, and winter storms cause deep mixing
down to 600 – 700 m. Light intensity during the day is low and day length is short, therefore the
critical depth (the depth where phytoplankton growth is equivalent to losses) is much shallower
than the mixing depth. During this season, respiration greatly exceeds photosynthetic input so
the result is a net loss of energy to the system. At the onset of spring, light increases in both
intensity and duration, thus driving the critical depth deeper. Storms begin to subside and the
increase of light combined with high concentrations of nutrients from previous mixing causes a
substantial photosynthetic gain. The spring bloom occurs when positive heat fluxes cause the
mixed layer to become shallower in the water column, thus allowing phytoplankton to remain
long enough above the critical depth to induce positive growth (Bagniewski et al., 2011). These
blooms directly influence the food available to deep-sea fauna through enhanced particle flux
(Herring, 2002). Organic particles are also consumed by microbes as they sink into the deep
ocean and therefore the speed in which they sink will determine the depth at which the particles
can be recycled (Bagniewski et al., 2011). In the summer, growth of the bloom is limited by the
5
rapid removal of nutrients. Losses due to grazing and sinking during this time are significant and
a corresponding increase of organic flocculent can be observed on the seafloor (Lampitt et al.,
2010, Martin et al., 2011). During autumn there is a decline in the number of zooplankton
grazers, and increased mixing from storms forces the mixed layer deeper, bringing nutrients to
the surface. The critical depth is relatively deep and light levels are still high, thus causing a
minor secondary phytoplankton bloom.
The fixation of carbon by phytoplankton and its transfer in food webs is strongly related
to the production of fished species (Nixon and Buckley, 2002). Previous estimates of capture
fisheries production in the Northeast Atlantic Ocean was about 8.6 million tons, ranking it the
most productive area in the Atlantic Ocean and fourth most productive when all areas of the
world’s oceans were considered (FAO, 2010). As coastal fisheries around the world become
depleted, commercial fishing has spread to the high seas and deep water in search of the last
economically attractive concentrations of fish biomass (Moore, 1999). Although the deep sea is
by far the largest ecosystem in the ocean, it is also the least productive and understood (Webb et
al., 2010). Topographic features such as seamounts and mid-ocean ridges can support
commercially valuable species because these features modify the physical and biological
dynamics in ways that enhance and retain food availability (Genin and Dower, 2007). Some
species form dense breeding aggregations over seamounts and mid-ocean ridges, therefore
increasing biomass concentrations and providing large catches. Most deep-sea fishes have life
history characteristics, such as slow growth and late maturity, which make them less resilient
than shallow-water fishes to heavy fishing pressure (Norse et al., 2012). It is therefore critical to
improve our understanding of the ecosystem processes that govern deep-sea fish population
dynamics in order to effectively manage vulnerable species.
6
Our knowledge of deep-sea ecosystems has increased dramatically since Forbes’ 1864
Azoic zone hypothesis, which stated that conditions in the deep sea are so unfavorable that it
should be devoid of life (Forbes, 1844). The first effort to examine the occurrence of deep-sea
fishes in the world’s oceans was the Challenger expedition in 1872, which ultimately disproved
the Azoic zone hypothesis (Thomson, 1873). Numerous species were described and, with the aid
of additional research expeditions, distribution patterns of biota were outlined. Since that time
many samples have been taken throughout the world’s oceans and we have accumulated enough
information to begin to interpret general distribution patterns, particularly for frequently sampled
organisms. These patterns can then be linked to hydrographic measurements to investigate
environmental drivers of species distributions. While the description of a species’ distribution is
critical, another important goal is to evaluate how distributions are maintained
trophodynamically. These gaps in knowledge were some of the major issues targeted by the
Census of Marine Life field project “Patterns and Processes of the Ecosystems of the Northern
Mid-Atlantic,” MAR-ECO.
The primary aim of MAR-ECO was to describe and understand the patterns of
distribution, abundance, and trophic relationships of the organisms inhabiting the mid-oceanic
North Atlantic, and to identify and model ecological processes that cause variability in these
patterns (Bergstad, O.A., Godø, 2003). Several research expeditions to the Mid-Atlantic Ridge,
from Iceland to the Azores, were conducted and results to date have demonstrated that ecosystem
processes along the MAR appear to operate differently from ‘typical’ open ocean ecosystems
(Sutton et al., 2008; Vecchione et al., 2010). One of the more surprising findings was a watercolumn maximum in fish biomass between 1500 and 2300 m (Sutton et al., 2008), which stands
in contrast to the typical view that fish biomass declines exponentially with increasing depth
7
from the surface in open-ocean ecosystems. Another ecologically significant finding was the
apparent influence of the Sub-Polar Front (SPF) on the distribution of various taxa along the
MAR. The SPF was the major hydrographic feature observed along the MAR and is
characterized by a near-surface frontal zone between cold, nutrient rich waters from the north
and warm, saline waters from the south. The ecological processes in this hydrographically
interesting area are not well understood and may influence the biogeographic and feeding
patterns of deep-sea fauna.
Reported catch numbers from MAR-ECO sampling indicated that Bathylagus euryops
(Argentiniformes: Microstomatidae) was the biomass dominant species along the MAR and
ranked third in total abundance (Sutton et al., 2008). Several studies have described the general
distribution and diet of B. euryops, however factors that cause variability within their distribution
and diet have not been investigated. In general, midwater fishes serve as an important link
between zooplankton and higher trophic level predators such as squids, piscivorous fishes,
seabirds, and marine mammals (Porteiro and Sutton, 2007). Although there are no current
fisheries for B. euryops, it is an important prey item for higher trophic levels, including
commercially important species such as orange roughy (Hoplostethus atlanticus), alfonsino
(Beryx splendens), and Greenland halibut (Reinharditus hippoglossoides) (Mauchline and
Gordon, 1984; Chambers and Dick, 2007; Horn et al., 2010). Given the high biomass and
abundance observed along the MAR combined with its importance as a prey item, B. euryops
appears to play an important role in the oceanic food web of the North Atlantic.
The lack of information available for most deep-sea fishes precludes the formulation of
proper management plans, thereby exacerbating the long-term effects of exploitation at current
levels of fishing (Cotton, 2010). In order to properly manage deep-sea fisheries, managers need
8
basic ecological information, including but not limited to relative species abundances,
distribution patterns, trophic interactions, feeding rates, the influence of ontogeny on distribution
and feeding, and the relationship of physical conditions to these biological patterns. In order to
address this need, I have provided detailed ecological information for B. euryops populations
along the MAR. In Chapter 1 of my thesis, I characterize the distribution of Bathylagus euryops
along the MAR and determine factors that influence this distribution. In Chapter 2, I describe
the feeding ecology of Bathylagus euryops along the MAR, including their diet, how physical
conditions affect their feeding, and calculate a daily ration. All of this information will be useful
in advancing our knowledge of the role of B. euryops in oceanic food webs. Ideally, this
information will be important in the establishment of proper management guidelines for
commercially important species that depend upon B. euryops as an abundant food source.
9
References
Angel, M.V., 1997. What is the deep sea? In: Randall, D.J., Farrell, A.P. (Eds.) Deep-Sea Fishes,
Academic Press, San Diego, pp. 1-41.
Bagniewski, W., Fennel, K., Perry, M.J., D’Asaro, E.A., 2011. Optimizing models of the North
Atlantic spring bloom using physical, chemical and bio-optical observations from a
Lagrangian float. Biogeosciences 8, 1291-1307.
Behrenfeld, M.J., Falkowski, P.G., 1997. Photosynthetic rates derived from satellite-based
chlorophyll concentration. Limnology and Oceanography 42(1):1-20.
Bergstad, O.A., Godø, O.R., 2003. The pilot project “Patterns and processes of the northern
Mid-Atlantic”: aims, strategy and status. Oceanologica Acta 25, 219-226.
Chambers, C.A., Dick, T.A., 2007. Using environmental variables to predict the structure of
deep-sea Arctic fish communities: Implications for food web construction. Arctic,
Antarctic, and Alpine Research 39:2-9.
Cotton, C.C., 2010. Age, growth and reproductive biology of deep-water Chondrichthyans.
VIMS dissertation.
FAO (Food and Agriculture Organization of the United Nations), 2010. The State of the World’s
Fisheries and Aquaculture, FAO Fisheries and Aquaculture Department, Rome, 2010.
Forbes, E., 1844. Report on the Mollusca and Radiata of the Aegean Sea, and on their
distribution, considered as bearing on geology. Report of the British Association for the
Advancement of Science for 1843, 129-193.
Genin, A., Dower, J.F., 2007. Seamount plankton dynamics. In: Pitcher, T.J., Morato, T., Hart,
P.J.B., Clark, M.R., Haggan, N., Santos, R.S. (Eds.), Seamounts: Ecology, Fisheries &
Conservation, Blackwell Publishing, Oxford, UK, pp. 85-100.
Haedrich, R.L., 1997. Distribution and population ecology. In: Randall, D.J., Farrell, A.P. (Eds.),
Deep-Sea Fishes, Academic Press, San Diego, CA, pp. 79-106.
Herring, P., 2002. The deep-sea dimension. In: Crawley, M.J., Little, C., Southwood, T.R.E.,
Ulfstrand, S. (Eds.), The Biology of the Deep Ocean, Oxford University Press, New
York, pp. 1-26.
Horn, P.L., Forman, J., Dunn, M.R., 2010. Feeding habits of alfonsino Beryx splendens. Journal
of Fisheries Biology 76:2382-2400.
Jennings, S., Kaiser, M.J., Reynolds, J.D., 2001. Marine fisheries ecology: an introduction. In:
Jennings, S., Kaiser, M.J., Reynolds, J.D (Eds.), Marine Fisheries Ecology, Blackwell
Publishing, Malden, MA, pp. 1-38.
10
Lampitt, R.S., Salter, I., de Cuevas, B.A., Hartman, S., Larkin, K.E., Pebody, C.A., 2010. Longterm variability of downward particle flux in the deep northeast Atlantic: Causes and
trends. Deep-Sea Research II 57:1346-1361.
Martin, P., Lampitt, R.S., Perry, M.J., Sanders, R., Lee, C., D’Asaro, 2011. Export and
mesopelagic particle flux during a North Atlantic spring diatom bloom. Deep-Sea
Research I 58:338-349.
Mauchline, J., Gordon, J.D.M., 1984. Occurrence and feeding of berycomorphid and
percomorphid teleost fish in the Rockall Trough. Journal du Counseil –Conseil
International pour l’exploration de la mer 41:239-247.
Moore, J.A., 1999. Deep-sea finfish fisheries: lessons from history. Fisheries 24(7):16-21
Nixon, S.W., Buckley, B.A., 2002. ‘A strikingly rich zone’ – nutrient enrichment and secondary
production in coastalmarine ecosystems. Estuaries 25:782-796.
Norse, E.A., Brooke, S., Cheung, W.W.L., Clark, M.R., Ekeland, I., Froese, R., Gjerde, K.M.,
Haedrich, R.L., Heppell, S.S., Morato, T., Morgan, L.E., Pauly, D., Sumaila, R., Watson,
R., 2012. Sustainability of deep-sea fisheries. Marine Policy 36:307-320.
Porteiro, F.M., Sutton, T., 2007. Midwater fish assemblages and seamounts. In: Pitcher, T.J.,
Morato, T., Hart, P.J.B., Clark, M.R., Haggan, N., Santos, R.S. (Eds.), Seamounts:
Ecology, Fisheries & Conservation, Blackwell Publishing, Oxford, UK, pp. 101-111.
Ramirez-Llodra, E., Brandt, A., Danovaro, R., De Mol, B., Escobar, E., German, C.R., Levin,
L.A., Martinez Arbizu, P., Menot, L., Buhl-Mortensen, P., Narayanaswamy, B.E., Smith,
C.R., Tittensor, D.P., Tyler, P.A., Vanreusel, A., Vecchione, M., 2010. Deep, diverse,
and definitely different: unique attributes of the world’s largest ecosystem.
Biogeosciences 7:2851-2899.
Sutton, T.T., Porteiro, F.M., Heino, M., Byrkjedal, I., Langhelle, G., Anderson, C.I.H., Horne, J.,
Søiland, H., Falkenhaug, T., Godø, O.R., Bergstad, O.A., 2008. Vertical structure,
biomass and topographic association of deep-pelagic fishes in relation to a mid-ocean
ridge system. Deep-Sea Research II 55:161-184.
Thomson, C.W., 1873. The Depths of the Sea. McMillan and Co., London.
Vecchione, M., Bergstad, O.A., Byrkjedal, I., Falkenhaug, T., Gebruk, A., Gislason, A., Heino,
M., Høines, Å.,Menezes, G., Piatkowski, U., Priede, I., Skov, H.,Søiland, H., Sutton, T.,
Wenneck, T. de Lange, 2010. Biodiversity patterns and processes on the Mid-Atlantic
Ridge. In: McIntyre, A. (Ed.), Life in the World’s Oceans: Diversity, Distribution, and
Abundance, Wiley-Blackwell, Oxford, UK, pp. 103-121.
11
Webb, T.J., Berghe, E.V., O’Dor, R., 2010. Biodiversity’s big wet secret: The global distribution
of marine biological records reveals chronic under-exploration of the deep pelagic ocean.
PLoS ONE 5(8):1-6.
12
CHAPTER 1
DISTRIBUTION OF BATHYLAGUS EURYOPS (TELEOSTEI: MICROSTOMATIDAE)
ALONG THE NORTHERN MID-ATLANTIC RIDGE
13
Abstract
The northern Mid-Atlantic Ridge (MAR), from Iceland to the Azores, ranges in depth
from 800 – 4500 m and extends over an area of 3.7 million km2. Despite its size, few studies
have described the distribution of pelagic fishes along the MAR. Recent evidence from MARECO, a Census of Marine Life field project, reported increased abundance and biomass of deeppelagic fishes below 1000 m on the ridge, which stands in stark contrast to the traditional view
that abundance and biomass decline exponentially with increasing depth in ‘typical’ open-ocean
ecosystems. Among the midwater fishes sampled during the MAR-ECO campaign, Bathylagus
euryops (Argentiniformes: Microstomatidae) was the biomass-dominant species and ranked third
in total abundance. In this paper, we characterize the distribution of B. euryops as a function of
physical and biological parameters along the MAR. General linear models were fitted to B.
euryops catch-per-unit-effort (CPUE) data to investigate possible influences on the distribution
of this species. Analyses of diel and ontogenetic vertical migrations were also conducted to
investigate the occurrence of B. euryops frequently observed in the epipelagic zone along the
MAR. Results indicated that a model containing categorical depth and geographic location
variables provided the most parsimonious description of B. euryops CPUE data, and that the
vertical distribution of this species is likely related to ontogeny. Due to the high abundance and
biomass observed along the MAR, B. euryops appears to be an important species in the oceanic
food web of the North Atlantic Ocean.
14
1. Introduction
The deep-pelagic realm, ranging from 200 m below the ocean’s surface to just above the
sea floor, is the largest habitat on Earth and holds 97% of the global ocean’s volume (Tyus,
2012). Generally thought of as one huge, continuous ecosystem, there is significant habitat
heterogeneity in the deep sea, as exemplified by mid-ocean ridges and seamounts, which can
have important impacts on ecosystem structuring and overall productivity. The Atlantic Ocean
contains one of the largest topographic features in the world, the Mid-Atlantic Ridge (MAR).
Despite the enormous size of the MAR, few studies have focused on characterizing the
distribution of pelagic fauna along it (Vecchione et al., 2010), which, in part, inspired the
initiation of the Census of Marine Life field project “Patterns and Processes of the Ecosystems of
the Northern Mid-Atlantic,” MAR-ECO (Bergstad and Godø, 2003; Bergstad et al., 2008).
Open ocean ecosystems are ultimately dependent upon near-surface productivity for
energy transfer to the meso- and bathypelagic zones. Food is extremely scarce in the deep
pelagial and must first pass through a suite of vertebrate, invertebrate and bacterial consumers in
the mesopelagic zone before reaching the bathypelagic zone. Therefore, there is a decrease in
production with increasing distance from the surface. Vertical distribution profiles in the North
Atlantic have shown that deep-pelagic zooplankton and micronekton biomass decline
exponentially with depth (Angel and Baker, 1982; Reinthaler et al., 2006; Arístegui et al., 2009;
Robinson et al., 2010). However, recent evidence of increased abundance and biomass of deeppelagic fishes below 1000 m on the MAR (Sutton et al., 2008) contradicts the traditional view of
vertical structuring in the open ocean. These observations suggest that ecosystems over the
MAR are structured differently than ‘typical’ open ocean ecosystems (Fock et al. 2004; Sutton et
al., 2008; Fock, 2009).
15
Among the fishes sampled during the 2004 MAR-ECO R/V G.O. Sars expedition,
Bathylagus euryops Goode and Bean 1896 was the biomass-dominant species and ranked third in
total abundance (Sutton et al., 2008). Historically, B. euryops is known to inhabit the North
Atlantic meso- and bathypelagic zones as far north as Greenland, extending south to Bermuda
(Mauchline and Gordon, 1983; Cohen, 1984; Møller et al., 2010) and is not known to migrate
vertically. However, Scott and Scott (1988) noted that B. euryops may be present in other
oceans, thus highlighting the uncertainty of even the most basic ecology of this species. The
deep-pelagic realm, overall, has been poorly sampled due to the technological challenges and
high cost of sampling (Webb et al., 2010). Therefore, knowledge of large-scale horizontal and
vertical distributions is limited for most deep-sea taxa, especially fauna in the vicinity of the
complex topographic features of the MAR (Vecchione et al., 2010).
A central theme in ocean ecology is the relationship between biological patterns and
regional physical conditions (Haedrich, 1997). In the North Atlantic Ocean, physical conditions
can vary greatly across a large-scale latitudinal gradient. The open ocean is also structured in the
vertical dimension, creating ecologically distinct habitats, and can be delineated by differences in
solar illumination, temperature, hydrostatic pressure, and current shear (Angel, 1986; Tyus,
2012). Within these vertically structured habitats, abrupt topographical features, such as
seamounts and mid-ocean ridges, can impinge on the deep-pelagic realm (Porteiro and Sutton,
2007). Major topographical features, such as the MAR, are known to influence local and
regional circulation patterns (Roden, 1987) and thus can have important implications on the
distribution of deep-pelagic organisms and their interactions with benthic fauna.
Along the MAR, Sutton et al. (2008) summarized the occurrence of bathypelagic fishes
in the epipelagic zone, including abundant species not known to migrate vertically such as
16
Cyclothone microdon and B. euryops. Physical factors may contribute to these previously
unobserved occurrences; however, behavioral or biological factors such as vertical migration
(both diel and ontogenetic) must also be considered. In this paper, we describe the distribution
of B. euryops over the MAR relative to hydrography, prey availability, ontogeny, and diel
patterns to understand better the influence each factor has on the habitat utilization of this
biomass-dominant species.
2. Methods
Specimens were collected during two separate MAR-ECO expeditions. The first was
Leg 1 of the R/V G.O. Sars expedition (Bergstad et al., 2008), 5 June – 3 July, 2004 from
Iceland to the Azores. Sampling occurred at 18 predefined stations along the MAR with
coverage across four different ridge sections (Fig. 1): Reykjanes Ridge (RR), Charlie-Gibbs
Fracture Zone (CGFZ), Faraday Seamount Zone (FSZ), and Azorean Zone (AZ). Vertical
profiles from the surface to near-bottom were generated with a CTD at all stations to characterize
the physical and chemical environment at the time of biological sampling (Søiland et al., 2008).
The CTD had dual sets of temperature and conductivity sensors, an Aquatracker III fluorometer
and a SeaBird oxygen sensor (SBE43). All sensors were calibrated prior to the research cruise.
The second expedition to the MAR was conducted aboard the NOAA FSV Henry B.
Bigelow, 12 June – 17 July, 2009. The objective of this expedition was to sample the fauna of
the near the Sub-Polar Front (SPF) region as previous studies indicated that the SPF influences
the faunal distribution of assemblages along the MAR (Hosia et al., 2008; Sutton et al., 2008;
Vecchione et al., 2010). Comparative day/night sampling occurred at 11 stations in two
transects, one with five stations northwest of the CGFZ, and one with six stations southeast of
17
the CGFZ (Fig. 2). Target depths were selected based on bottom depth with the goal to
standardize the depth zones sampled to match those on the G.O. Sars cruise.
2.1. Trawl Sampling
Samples for this study were collected using a double-warp midwater ‘Krill’ trawl (6×6 m
mouth opening, 6-mm mesh, stretched) deployed with standard pelagic trawl doors. The total
trawl length from mouth to cod end was 45 m and the net was equipped with a multi-sampler
(remotely operated multiple cod ends) with five 30-m long cod ends that enabled five discretedepth samples per tow. Each cod end was rigged with a 7-L hard collection receptacle, in order
to reduce physical damage to the organisms sampled. Additionally, the trawl was equipped with
SCANMAR sensors to provide data on net mouth configuration and depth.
Sampling was performed along an oblique trajectory, from 3000 m to the surface, and
nets were remotely opened and closed based on command from the ship during the 2004
expedition and a programmed timer was used during the 2009 expedition. The trawl was towed
at an average speed of 2.2 knots and the duration that each depth stratum was sampled ranged
between 10-40 minutes. The volume of water filtered was subsequently calculated for each cod
end based on the door spread distance, speed of the ship during each sampling event, and
duration of each sampling event. Further methodological details can be found in Wenneck et al.
(2008).
In total, 113 discrete-depth samples were collected during the two research expeditions
and used to characterize the distribution of B. euryops along the MAR. Samples were typically
collected within one of five depth zones: (1) 0-200 m, (2) 200-750 m, (3) 750-1500 m, (4) 15002300 m and (5) >2300 m. Time of day was classified as day (D) or night (N) for each sample
18
based sunrise and sunset times for each sampling location. Samples that were collected during
‘twilight’ hours (one hour before and after sunrise/sunset), were not included in analyses where
time of day was concerned. Following trawl retrieval, the catch was sorted by individual cod
ends, corresponding to the depth stratum sampled, into major taxonomic groups which were
further identified to species level.
2.2. Sample handling
During the 2004 G.O. Sars expedition, species abundance and total biomass (± 0.1 g)
were recorded for each cod end and specimens were frozen at -20 ⁰C at sea. When taxonomic
identity was uncertain, specimens were preserved in 10% formalin and later identified at the
Bergen Museum of Zoology, http://collections.uib.no/vertebrate/, the permanent repository for
all pelagic fishes sampled during the 2004 expedition. During the 2009 Henry B. Bigelow
expedition, B. euryops specimens were enumerated, measured (standard length, ± 0.1 mm),
weighed individually (wet weight, ± 0.1 g), and frozen at -80 ⁰C at sea. All pelagic fishes
collected during this expedition were then stored at the Virginia Institute of Marine Science.
2.3. Statistical analyses
Fish abundance data collected during the 2004 G.O. Sars cruise were standardized into
catch-per-unit-effort (CPUE) (individuals •10-3·m-3 water filtered) and analyzed using general
linear models (GLM). A GLM is defined by the underlying statistical distribution for the
response variable and how a set of linearly related explanatory variables correspond to the
expected value of the response variable (Maunder and Punt, 2003). The linear relationship
19
between explanatory variables and some function of the expected value of the response variable
is given by:
g ( µ i ) = x iT β ,
(1)
where µi = E(Yi) with Yi defined as the ith value of the response variable, xi is the vector of
associated explanatory variables, β is a vector of estimated parameters, and g is the differentiable
monotonic link function that governs the relationship (linear or nonlinear) between the random
(µi) and systematic ( β
x iT ) components of the model (McCullagh and Nelder, 1989; Maunder and
Punt, 2003).
Histograms of raw CPUE data by ridge section and depth zone were consistently
positively skewed which provided some evidence that the CPUE data may be lognormally
distributed. Visual examination of QQ plots of normality and residuals plots for preliminary
model fits assuming normally and lognormally distributed CPUE data provided additional
support for application of a lognormal GLM (Zuur et al., 2010). Given the need for log
transformation, a constant of 10-5 was added to all CPUE values to adjust for zero catches while
maintaining the relative scale and the distributional shape of the data (note that model results
were generally insensitive to the value of the constant added).
In an effort to identify the most parsimonious description of the CPUE data, seven model
parameterizations were fitted where each reflected a unique hypothesis about the effects of
various combinations of explanatory variables: M1 included depth zone, M2 contained ridge
section, M3 included both depth zone and ridge section, M4 contained depth zone, ridge section,
and the depth*ridge interaction, M5 included depth zone, ridge section, and prey abundance
20
(derived from Amphipoda, Decapoda, Euphausiacea, and Cnidaria data collected with ‘Krill’
trawl during cruises that were potential prey types for B. euryops; Vecchione et al., unpublished
data), M6 included depth zone, ridge section, prey abundance, and the depth*prey interaction,
and M7 included depth zone, ridge section, prey abundance, depth*prey, and depth*ridge
interactions. Other explanatory variables, including temperature, salinity, and depth from
bottom, were also investigated during preliminary analyses of the data, but were ultimately not
included in the parameterization analysis due to lack of significance. Using model M5 as an
example (full main effects model), the exact structure is given by:
log e (CPUE + 10 −5 ) drtp = β 0 + β d xd + β r xr + β t xt + β p x p + ε drtp ,
(2)
where d, r, t, p index the depth zone (five levels), ridge section (four levels), and prey abundance
variables, respectively, and εdrtp are the error terms. The identity link function was applied for all
fitted models.
Model comparisons were made using Akaike’s Information Criterion corrected for smallsample size (AICc):
2k (k + 1)
,
AIC c = −2 log( Lˆ ) + 2k +
n − k −1
(3)
where L̂ is the estimated maximum likelihood value, k is the number of model parameters, and n
is the total sample size. The most parsimonious model of a given set has the lowest AICc value,
and because AICc is on a relative scale, it is important to calculate AIC differences (∆AIC):
21
∆AIC c = AIC c − AIC c min ,
(4)
where AICc min is the smallest AICc value within the candidate set of models. Generally, ∆AIC
values between 0 and 2 are indicative of substantial empirical support for the fitted model, values
between 4 and 7 are associated with models that have considerably less empirical support, and
values >10 suggest virtually no empirical support (Burnham and Anderson, 2002). Predicted
mean CPUEs were generated from the best fitting model by back transforming and multiplying
by the bias correction term of exp(0.5σˆ 2 ) , where σˆ 2 is the estimated residual error. Standard
errors of the predicted mean CPUEs were estimated using a standard jackknife routine (Efron,
1981).
The occurrence of B. euryops in the epipelagic zone along the MAR is atypical for the
species. Utilizing CPUE data from the 2009 FSV Henry B. Bigelow expedition, the possibility of
B. euryops undergoing a diel vertical migration was investigated. Again, a log transformation
was required to meet the assumptions of analysis with linear models, so a constant of 10-5 was
added to all CPUE values to adjust for zero values associated with the catch data. The influence
of time of day on the vertical distribution of B. euryops was examined using a general linear
model of the following structure:
log e (CPUE + 10 −5 ) ds = β 0 + β d x d + β s x s + ε ds
where d and s index the depth zone (five levels) and time of day (two levels) variables,
respectively, and εds are the error terms. The identity link function was again applied.
22
(5)
Changes in vertical distribution as a result of ontogenetic habitat shifts could also explain
the occurrence of B. euryops in surface waters near the Charlie-Gibbs Fracture Zone. Another
general linear model was utilized to investigate this possibility with data from the Henry B.
Bigelow expedition, where fish size (standard length, mm) was modeled as a function of depth to
further elucidate patterns in the vertical distribution of B. euryops. QQ plots of normality and
residuals plots for preliminary model fit assuming normally distributed fish size data were
visually examined and demonstrated a normal distribution of the response variable and a
reasonable model fit. The exact structure of the linear model with the identity link function is as
follows:
Fish Size d = β 0 + β d x d + ε d
(6)
where d indexes the depth zone (five levels) variable and εd are the error terms.
Analyses of variance (ANOVA) was applied to the best fitting general linear model to
facilitate post-hoc comparisons (pairwise t-tests) of mean CPUEs or fish sizes within treatment
levels (e.g. depth, ridge section). A Bonferroni adjustment was applied to the pairwise t-tests to
control for Type I error (α = 0.05).
3. Results
Bathylagus euryops represented ~8% (n=4,543) of the total fish catch and ~25% (190,027
g ww) of the total biomass sampled during the 2004 G.O. Sars expedition (Sutton et al., 2008).
During the 2009 Henry B. Bigelow expedition, B. euryops were < 2% of the catch (n=243), but
contributed ~20% of the total pelagic fish biomass (6,436 g ww). Individual fish length were not
23
recorded during the G.O. Sars expedition, but ranged from 24.0 – 229.1 mm during the Henry B.
Bigelow expedition.
Composite remote-sensing images identified the position of the Sub-Polar Front (SPF),
which was near 52 ⁰N, although the southern boundary of the SPF was very indistinct (Søiland et
al., 2008). Additionally, four different hydrographic regions were identified in the upper 500 m
along the MAR. Modified North Atlantic Water dominated at all stations north of 57 ⁰N on the
Reykjanes Ridge. Stations south of 57 ⁰N, but north of the SPF, were dominated by Sub Arctic
Intermediate Water. South of the SPF, a mixture of North Atlantic Modified Water and Sub
Arctic Intermediate Water was observed and formed a complex pattern of eddies to 50 ⁰N. At all
stations south of 50 ⁰N, the predominant surface water mass was North Atlantic Central Water
(Vecchione et al., 2010).
Based on AICc and thus ∆AICc, model M3 received the most empirical support, followed
closely by model M5. There was considerably less support for models M4 and M6 and virtually
no support for models M1, M2, and M7 (Table 1). Model M3 contained depth and ridge section as
explanatory variables and all of the estimated βi coefficients associated with the various levels of
those categorical variables, except that for depth zone 5 (> 2300 m), were statistically significant
(Table 2). The model intercept, β0, is the reference level of the depth and ridge section variables
and it was defined to be the epipelagic zone in the AZ. Relative to the intercept estimate, all
other estimated coefficients were positive and fairly large in magnitude, which suggests that the
estimated CPUE of B. euryops is higher at all other depths and ridge sections sampled. The
highest estimated CPUEs occurred in depth zone 3 (750 – 1500 m) and at the CGFZ, with the
next highest estimated CPUEs being within depth zone 4 (1500 – 2300 m) and at the RR (note
that these were similar in magnitude to the highest). Prey availability indices, which included
24
various genera of amphipods, decapod shrimp, euphausiids and coronate jellyfish, were not
useful in explaining the distribution of B. euryops along the MAR.
ANOVA results based on M3 revealed that both ridge section (F-value: 14.07, p < 0.05)
and depth (F-value: 9.35, p < 0.05) had a significant effect on the mean CPUE of B. euryops.
Post-hoc multiple comparisons applied to the four sections highlighted significant differences in
the mean CPUE of B. euryops between the AZ and the two northern ridge sections (RR, p < 0.05
and CGFZ, p < 0.05). Mean CPUE differed weakly between the AZ and the FSZ (p = 0.106) and
all other ridge section multiple comparisons were not significant. Similar multiple comparisons
applied to the five depth strata resulted in significant differences in the mean CPUE of B.
euryops in the epipelagic zone when compared to the mean CPUE at depth zones 2, 3, and 4 (p <
0.05). No significant differences were detected between the epipelagic zone and depths zone 5
(p = 1.00) and all other depth strata multiple comparisons also demonstrated no significant
differences.
The diel vertical migration model contained solar cycle and depth zone as explanatory
variables, and of the estimated βi coefficients associated with the levels of those variables, only
depth zone 3, was statistically significant (Table 3). The intercept reference level, β0, was
defined to be the epipelagic zone during the day. Relative to the intercept estimate, the estimated
coefficient for night was negative, which suggests that CPUEs observed during the night were
lower than those observed during the day. Concurrent day/night sampling during the Henry B.
Bigelow expedition revealed that solar cycle did not significantly influence the vertical
distribution of B. euryops at the CGFZ (F-value: 0.15, p = 0.71).
The ontogenetic model, with fish size as the response variable, contained depth as the
only explanatory variable. Of the estimated βi coefficients associated with the various levels of
25
the categorical depth variable, depth zones 3 (750 – 1500 m), 4 (1500 – 2300 m), and 5 (> 2300
m) were statistically significant (Table 4). The model intercept, β0, is the reference level of the
depth variable and was defined to be the epipelagic zone. All estimated coefficients were
positive relative to the intercept estimate, meaning that there was a consistent increase in fish
size with increasing depth from the surface. As such, the lowest estimated mean standard length
occurred in surface waters whereas the highest estimated mean standard length occurred in the
deepest depth zone. ANOVA results of the ontogenetic model further illustrated a significant
relationship between fish size and depth (F = 6.93, p < 0.05) and predicted means relative to
each depth zone were plotted in Figure 4. Multiple comparisons statistics revealed that the mean
size of B. euryops were significantly smaller in the epipelagic zone than in depth zone 4 (1500 –
2300 m; p < 0.05) and depth zone 5 (> 2300 m; p < 0.05). Similarly, fish size was significantly
larger in depth zone 4 (1500 – 2300 m) when compared to those in depth zone 2 (200 – 750 m; p
< 0.05) and depth zone 3 (750 – 1500 m; p < 0.05).
4. Discussion
4.1. Along-ridge distribution patterns
The analyses performed on the discrete-depth trawl samples collected during the MARECO campaign represent the largest-scale geographical and vertical examination of the
distribution of B. euryops to date. Results from the linear model analyses of CPUE revealed that
a linear model containing depth and ridge section as explanatory variables (M3) provided the
most empirical support and that depth and ridge section were significant factors in the
distribution of B. euryops along the MAR. In general, the vertical distribution patterns of B.
euryops observed on the MAR agree with reported occurrences elsewhere in the North Atlantic.
26
Previous studies have reported B. euryops occurrences between 500 and 3000 m depth (Cohen,
1964; Mauchline and Gordon, 1983; Moore, 2003), indicating a degree of depth structuring for
this species. The fact that a deep-sea species’ distribution is structured, in part, by depth is not
too surprising as vertical zonation of pelagic fishes has been observed in all parts of the world’s
oceans (Haedrich and Merrett, 1988). However, the view of fishes zoned by depth is far too
simplistic considering depth is a multivariate factor. Instead, deep-sea fish distributions are more
likely linked to interacting physical gradients that act in concert with increasing depth from the
surface (Angel, 1986), with the effects of individual physical elements being hardly separable
from one another (Backus, 1986).
A clear geographic pattern was observed in the distribution of B. euryops, with maximum
total integrated abundances occurring at the SPF/Charlie-Gibbs Fracture Zone area and declining
southward to the Azorean Zone. Sutton et al. (2008) observed that the strongest differences in
overall fish abundance and biomass were between the Azorean Zone and the two northern ridge
sections (RR and CGFZ), which agrees with findings from this study (Fig. 3). Overall
abundance of B. euryops was lowest at the Azorean Zone, where nearly all fish were sampled
below 1500 m. The presence of B. euryops throughout the entire water column in the higher
latitudes combined with the exclusion of B. euryops from the mesopelagic zone in the lower
latitudes exemplifies the principles of tropical submersion, which has also been observed in other
deep-sea fishes such as ceratioid anglerfishes (Helfman et al., 2009). However, additional
sampling south of the Azores would be necessary to corroborate the concept of tropical
submersion and a subsequent extension of B. euryops distributions into the South Atlantic
Ocean. Alternatively, the Azorean Zone may be approaching the faunal boundaries of B.
euryops. Vecchione et al. (2010) noted that along-ridge abundance patterns appear associated
27
with water-mass distributions and circulation features, however depth and topography likely play
additional major structuring roles. Although there appeared to be an association of B. euryops
with the Sub-Polar Front region, our observations are inferred from a small snapshot in time.
The SPF varies both spatially (48-53⁰N) and temporally, and hence more information is needed
to understand how the distribution of B. euryops varies in time relative to hydrographic frontal
zones.
Findings from MAR-ECO have identified the SPF/Charlie-Gibbs Fracture Zone area as a
region of maximum chlorophyll a concentrations (Gaard et al., 2008), where along-ridge
abundance maximums were observed for a variety of taxa, including Calanus (Gislason et al.,
2008) and Paraeuchaeta copepods (Falkenhaug et al, 2007), chaetognaths (Pierrot-Bults, 2008),
gelatinous zooplankton (Youngbluth et al., 2008), and based on results from this study, B.
euryops as well. Food supply is a major factor that influences the distribution and abundance of
deep-sea fishes (Tyus, 2012), and all of the taxa listed above are known prey types in the diet of
B. euryops (Mauchline and Gordon, 1983). Given the overall scarcity of food in the deep-sea,
Genin (2004) postulated a trophodynamic mechanism, termed the ‘topographic blockage’
hypothesis, which suggests that seamounts and mid-ocean ridges are able to support enhanced
standing stocks of fishes. The mechanism is related to the daily ‘gap formation’ of zooplankton
over abrupt topographic features (Genin et al., 1994). During the night, zooplankton are
advected over the summits of seamounts by near-surface currents. At dawn, the descent of the
vertically migrating zooplankton is blocked by the shallow bottom and these areas typically
correlated with high abundance of predators. The predators then feed on the trapped
zooplankton, significantly reducing their numbers, and setting the stage for a gap formation
where few zooplankton remain to ascend to the seamount summit the following evening.
28
Although this mechanism would most benefit site-attached demersal fishes, it may also have key
implications for pelagic zooplanktivorous fishes like B. euryops. Wilson (1992) demonstrated
with a time series of acoustic transects that sound scattering micronekton and fishes were able to
maintain their position above the SE Hancock seamount despite strong currents that should have
displaced the organisms to the surrounding deep waters. The ability to maintain location on the
seamount, coupled with the observed year-long increase in body size of an apparent cohort of
Maurolicus muelleri (Boehlert et al, 1994), indicated that the aggregations of micronekton and
fish over the SE Hancock seamount were resident (Genin, 2004). Pelagic fishes would most
certainly need to expend more energy relative to demersal fishes to stay near structure. For the
‘topographic blockage’ mechanism to hold true for pelagic fishes, the energy lost in maintaining
their position against strong currents would need to be compensated by the enhanced trophic
subsidy to these fishes. However, the prey availability index used in this study was not useful in
explaining the distribution of B. euryops, likely due to gear constraints which lead to an
inadequate representation of the prey field as it relates to B. euryops. Specifically, the mesh size
of the ‘Krill’ trawl (6 mm, stretched) likely under-sampled a majority of the smaller
mesozooplankton (2-5 mm) that B. euryops primarily consumes. Furthermore, the physical act
of net sampling often destroys delicate gelatinous zooplankton (Harbison, 1986), an important
prey group of B. euryops, and thus more accurate abundance estimates are needed to fully
explore the role of prey availability in structuring the distribution of B. euryops.
4.2. Diel vertical migration
One unusual finding in this study was the presence of B. euryops in the epipelagic zone
along the MAR, occasionally in large numbers. This either represents a new vertical range
29
extension for the species, or the multi-sampler attached to the ‘Krill’ trawl was prone to selective
contamination. Sutton et al. (2008) argued that the occurrence of bathypelagic fish in the upper
200 m was not artifactual during the 2004 G.O. Sars expedition. During the 2009 Henry B.
Bigelow expedition, the observation of a variety of bathypelagic fauna, including B. euryops, in
the epipelagic zone, led to post-hoc sampling of surface waters in an effort to investigate the
contamination question. Consequently, catch observations from the post-hoc sampling effort of
the epipelagic zone revealed that the presence of numerous bathypelagic taxa in surface waters
was a biological signal and not an artifact of selective contamination. The occurrence of deeppelagic fishes in surface waters is yet another indication that the abundance patterns and
distribution of deep-pelagic fishes over the MAR is different from a ‘typical’ open-ocean
ecosystem. Analysis of catch data based on day/night sampling revealed that the presence of B.
euryops in the epipelagic zone was probably not due to an active diel vertical migration. From a
physiological perspective, fishes with lipid-filled swimbladders are capable of extensive vertical
migrations, whereas those with gas-filled swimbladders typically undergo reduced vertical
migrations due to difficulties in secreting gas at very high pressures (Herring, 2002). The lack of
a swimbladder in its entirety in B. euryops, coupled with greatly reduced musculature, would
make an active diel vertical migration energetically demanding. Thus, it is not likely that solar
cycle played a significant role in the vertical distribution of B. euryops.
4.3. Ontogenetic migration
In most, if not all midwater oceanic fishes, early development occurs in near surface
waters followed by a descent into deeper zones (Pietsch, 2009). However, like most deep-sea
fishes, early life history information is generally lacking for B. euryops (Fahay, 2007) and this
30
represents a critical knowledge gap in our understanding of its distribution. Results from this
study revealed a significant positive relationship between fish size and depth, with smaller fish
occurring in surface waters and larger individuals occurring at greater depth. Eggs of B. euryops
have multiple oil globules (Ahlstrom et al., 1984) and larvae are elongate, with large fin folds
and stalked eyes (Kenaley and Hamilton, Jr., 2006). These physiological traits are characteristic
of an early life history adapted to conditions in surface waters. Oil globules in eggs are less
dense than the surrounding water, thus providing lift and causing the eggs to ascend (Balon,
1975). The larval form of B. euryops optimizes its surface to volume ratio with temporary
characters (e.g. large fin folds, stalked eyes) that reduce the tendency to sink, since drag is
directly proportional to the surface area of an object (Pelster, 1997). By reducing the tendency to
sink, the ability of larval B. euryops to locate food and grow quickly is enhanced in the
productive waters of the euphotic zone. Findings from this study support the few, documented
early life history patterns of B. euryops, indicating that a gradual ontogenetic migration occurs
from the surface down to the bathypelagic realm. Thus, the presence of B. euryops in the
epipelagic zone along the MAR is likely a function of early life history, and the overall range of
the vertical distribution of B. euryops should subsequently be extended into surface waters.
5. Conclusion
The analyses of the distribution of B. euryops presented in this study as a function of life
history characteristics and environmental variables provides valuable insights into the structure
of deep-sea ecosystems along the understudied MAR. However, the task of determining
distribution patterns in an ecosystem as large and deep as the MAR requires an integrated
sampling approach, including obtaining information on ontogenetic and spatial changes in
31
distribution. The patterns described in the present study are merely snapshots from the summer,
after the peak in primary production from the spring bloom had already occurred. The primary
remaining gaps in the knowledge of distribution patterns of B. euryops along the MAR is the
determination of temporal patterns (seasonal and interannual) and resolving the influence of
hydrography and water masses.
Acknowledgements
The authors acknowledge the crew of the R/V G.O. Sars and the FSV Henry B. Bigelow
for their excellent shiptime services. Thanks are due to E. Hilton, D. Steinberg, and K. Parsons
for their valuable comments on the manuscript. The data analysis and synthesis were supported
by NSF Ocean Sciences Division – Biological Oceanography Program (OCE 0623551) and the
VIMS Office of Academic Studies. MAR-ECO is a Census of Marine Life field project.
32
References
Ahlstrom, E.H., Moser, H.G., Cohen, D.M., 1984. In: Argentinoidei: Development and
Relationships. In: Moser, H.G., Richards, W.J., Cohen, D.M., Fahay, M.P., Kendall,
A.W., Richardson, S.L. (Eds.), Ontogeny and Systematics of Fishes American Society of
Ichthyology and Herpetology Special Publication 1, pp.155-169.
Angel, M.V., Baker, A., 1982. Vertical distribution of the standing crop of plankton and
micronekton at three stations in the Northeast Atlantic. Biological Oceanography 2,
1-30.
Angel, M.V., 1986. Vertical distribution: Study and implications. In: Pierrot-Bults, A.C., van
der Spoel, S., Zahuranec, B.J., Johnson, R.K. (Eds.), Pelagic Biogeography: Proceedings
of an International Conference, The Netherlands 29 May-5 June 1985, Paris, UNESCO
49, pp. 3-8.
Arístegui, J., Gasol, J.M., Duarte, C.M., Herndl, G.J., 2009. Microbial oceanography in the dark
ocean’s pelagic realm. Limnology and Oceanography 54(5), 1501–1529.
Backus, R.H., 1986. Biogeographic boundaries in the open ocean. In: Pierrot-Bults, A.C., van
der Spoel, S., Zahuranec, B.J., Johnson, R.K. (Eds.), Pelagic Biogeography: Proceedings
of an International Conference, The Netherlands 29 May-5 June 1985, Paris, UNESCO
49, pp. 9-13.
Balon, E.K., 1975. Reproductive guilds of fishes: A proposal and definition. Journal of the
Fisheries Board of Canada 32, 821-864.
Bergstad, O.A., Godø, O.R., 2003. The pilot project “Patterns and processes of the northern
Mid-Atlantic”: aims, strategy and status. Oceanologica Acta 25, 219-226.
Bergstad, O.A., Falkenhaug, T., Astthorsson, O.S., Brykjedal, I., Gebruk, A.V., Piatkowski, U.,
Priede, I.G, Santos, R.S., Vecchione, M., Lorance, P., Gordon, J.D.M., 2008. Towards
improved understanding of the diversity and abundance patterns of the mid-ocean ridge
macro- and megafauna. Deep-Sea Research II 55, 1-5.
Boehlert, G.W., Wilson, C.D., Mizuno, K., 1994. Populations of the sternoptychid fish
Maurolicus muelleri on seamounts in Central North Pacific. Pacific Science 1, 57– 69.
Burnam, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference: A Practical
Information-Theoretical Approach 2nd ed., Springer-Verlag, NewYork, NY.
Cohen, D.M., 1964. Bathylagidae. In: Bigelow, H.B., Breder, C.M., Cohen, D.M., Mead, G.W.,
Merriman, D., Olsen, Y.H., Schroeder, W.C., Schultz, Tee-Van, J. (Eds.), Fishes of the
Western North Atlantic, Sears Foundation for Marine Research, New Haven, CT, pp. 3448
33
Cohen, D.M., 1984. Bathylagidae. In: Whitehead, P.J.P., Bauchot, M.L., Hureau, J.C., Nielsen,
J., Tortonese, E. (Eds.), Fishes of the North-Eastern Atlantic and Mediterranean, vol. 1,
UNESCO, Paris, pp. 392–394.
Efron, B., 1981. Nonparametric estimates of standard error: The jackknife, the bootstrap and
other methods. Biometrika 68, 589-599.
Fahay, M.P., 2007. Bathylagidae. In: Early Stages of Fishes in the Western North Atlantic
Ocean., vol. 1, Northwest Atlantic Fisheries Organization, Dartmouth, Nova Scotia, pp.
230-231.
Falkenhaug, T, Gislason, A., Gaard, E., 2007. Vertical distribution and population structure of
copepods along the northern Mid-Atlantic Ridge. ICES ASC Helsinki 17-21 September
2007, ICES CM 2007/F:07.
Fock, H.O., Pusch, C., Ehrich, S., 2004. Structure of deep-sea pelagic fish assemblages in
relation to the Mid-Atlantic Ridge (45°-50° N). Deep-Sea Research 51, 953-978.
Fock, H.O., 2009. Deep-sea pelagic ichthyonekton diversity in the Atlantic Ocean and the
adjacent sector of the Southern Ocean. Global Ecology and Biogeography 18, 178-191.
Gaard, E., Gislason, A., Falkenhaug, T., Søiland, H., Musaeva, E., Vereshchaka, A., Vinogradov,
G., 2008. Horizontal and vertical copepod distribution and abundance on the MidAtlantic Ridge in June 2004. Deep-Sea Research II 55, 59-71.
Genin, A., Greene, C., Haury, L., Wiebe, P., Gal, G., Kaartvedt, S., Meir, E., Fey, C., Dawson,
J., 1994. Zooplankton patch dynamics: daily gap formation over abrupt topography.
Deep-Sea Research I 41, 941-951.
Genin, A., 2004. Bio-physical coupling in the formation of zooplankton and fish aggregations
over abrupt topographies. Journal of Marine Systems 50, 3-20.
Gislason, A., Gaard, E., Debes, H., Falkenhaug, T., 2008. Abundance, feeding and reproduction
of Calanus finmarchicus in the Irminger Sea and on the northern Mid-Atlantic Ridge in
June. Deep-Sea Research II 55, 72-82.
Haedrich, R.L, Merrett, N.R., 1988. Summary atlas of deep-living demersal fishes in the North
Atlantic basin. Journal of Natural History 22, 1325–1362.
Haedrich, R.L., 1996. Deep-water fishes: Evolution and adaptation in the earth’s largest living
spaces. Journal of Fish Biology 49, 40-53.
Haedrich, R.L., 1997. Distribution and Population Ecology. In: Randall, D.J., Farrell, A.P.
(Eds.), Deep-Sea Fishes, Academic Press, San Diego, CA, pp. 79-106.
Harbison, G.R., 1986. Vertical distribution: Study and implications. In: Pierrot-Bults, A.C., van
der Spoel, S., Zahuranec, B.J., Johnson, R.K. (Eds.), Pelagic Biogeography: Proceedings
34
of an International Conference, The Netherlands 29 May-5 June 1985, Paris, UNESCO
49, pp. 112-117.
Hartel, K.E., Kenaley, C.P., Galbraith, J.K., Sutton, T.T., 2008. Additional records of deep-sea
fishes from off greater New England. Northeast Naturalist 15(3), 317-334.
Helfman, G.S., Collette, B.B., Facey, D.E., Bowen, B.W., 2009. The Diversity of Fishes:
Biology, Evolution, and Ecology, Wiley-Blackwell, West Sussex, UK.
Herrring, P., 2002. On being efficient. In: Crawley, M.J., Little, C., Southwood, T.R.E.,
Ulfstrand, S. (Eds.), The Biology of the Deep Ocean, Oxford University Press, New
York, NY, pp. 111-117.
Hosia, A., Stemmann, L., Youngbluth, M., 2008. Distribution of net-collected planktonic
cnidarians along the northern Mid-Atlantic Ridge and their associations with the main
water masses. Deep-Sea Research II 55, 106-118.
Kenaley, C.P., Hamilton, Jr., A.N., 2006. Bathylaginae: Deep-sea Smelts. In: Richards, W.J.
(Ed.), Early Stages of Atlantic Fishes: An Identification Guide for the Western Central
North Atlantic vol. 1, CRC Press, Boca Raton, FL, pp. 141-151.
Mauchline, J., Gordon, J.D.M., 1983. Diets of clupeoid, stomiatoid, and salmonoid fish of the
Rockall Trough, northeastern Atlantic Ocean. Marine Biology 77, 67-78.
Maunder, M.N., Punt, A.E., 2004. Standardizing catch and effort data: a review of recent
approaches. Fisheries Research 70, 141–159.
McCullagh, P., Nelder, J.A., 1989. Generalized Linear Models, CRC Press, Boca Raton, FL, pp.
8-16.
Møller, P.R., Nielsen, J.G., Knudsen, S.W., Poulsen, J.Y., Sünksen, K., Jørgensen, O.A., 2010.
A checklist of the fish fauna of Greenland waters. Zootaxa 2378, 23.
Moore, J.A., Vecchione, M., Hartel, K.E., Collette, B.B., Galbraith, J.K., Gibbons, R.,
Turnipseed, M., Southworth, M., Watkins, E., 2003. Biodiversity of Bear Seamount,
New Elngland seamount chain: results of exploratory trawling. Journal of Northwest
Atlantic Fisheries Science 31, 363-372.
Moyle, P.B., Cech, Jr., J.J., 2000. Deepsea Habitats. In: Ryu, T. (Ed.), Fishes: An Introduction
to Ichthyology, Prentice Hall, Upper Saddle River, NJ, pp. 490-492.
Pelster, B., 1997. Buoyancy at Depth. In: Randall, D.J., Farrell, A.P. (Eds.), Deep-Sea Fishes,
Academic Press, San Diego, CA, pp. 198.
Pierrot-Bults, A.C., 2008. A short note biogeographic patterns of the Chaetognatha fauna in the
North Atlantic. Deep-Sea Research II 55, 137-141.
35
Pietsch, T.W., 2009. Oceanic Anglerfishes: Extraordinary Diversity in the Deep Sea, University
of California Press.
Porteiro, F.M., Sutton, T., 2007. Midwater fish assemblages and seamounts. In: Pitcher, T.J.,
Morato, T., Hart, P.J.B., Clark, M.R., Haggan, N., Santos, R.S. (Eds.), Seamounts:
Ecology, Fisheries & Conservation, Blackwell Publishing, Oxford, UK, pp. 101-111.
Reinthaler, T., van Aken, H., Veth, C., Arístegui, J., Robinson, C., Williams, P., Lebaron, P.,
Herndl, G.J., 2006. Prokaryotic respiration and production in the meso- and bathypelagic
realm of the eastern and western North Atlantic basin. Limnology and Oceanography
51(3), 1262–1273.
Robinson, C., Steinberg, D.K., Andersen, T.R., Arístegui, J., Carlson, C.A., Frost, J.R.,
Ghiglione, J.F., Hernández-Léon, S., Jackson, G.A., Koppelmann, R., Quéguiner, B.,
Ragueneau, O., Rassoulzadegan, F., Robison, B.H., Tamburini, C., Tanaka, T., Wishner,
K.F., Zhang, J., 2010. Mesopelagic zone ecology and biogeochemistry – a synthesis.
Deep-Sea Research II 57, 1504-1518.
Roden, G.I., 1987. Effect of seamount chains on ocean circulation and thermohaline structure.
In: Boehlert, G.W. (Ed.), Geophysical Monograph 43. American Geophysical Union,
Washington, DC, pp. 335-354.
Scott, W.B., Scott, M.G., 1988. Atlantic fishes of Canada. Canadian Bulletin of Fisheries and
Aquatic Sciences 219, 159-160.
Søiland, H., Budgell, P., Knutsen, Ø., 2008. The physical oceanographic conditions along the
Mid Atlantic Ridge north of the Azores in June–July 2004. Deep-Sea Research II 55, 2944.
Sutton, T.T., Porteiro, F.M., Heino, M., Byrkjedal, I., Langhelle, G., Anderson, C.I.H., Horne, J.,
Søiland, H., Falkenhaug, T., Godø, O.R., Bergstad, O.A., 2008. Vertical structure,
biomass and topographic association of deep-pelagic fishes in relation to a mid-ocean
ridge system. Deep-Sea Research II 55, 161-184.
Tyus, H.M., 2012. Ecology and Conservation of Fishes, CRC Press, Boca Raton, FL.
Vecchione, M., Bergstad, O.A., Byrkjedal, I., Falkenhaug, T., Gebruk, A., Gislason, A., Heino,
M., Høines, Å., Menezes, G., Piatkowski, U., Priede, I., Skov, H., Søiland, H., Sutton, T.,
Wenneck, T. de Lange, 2010. Biodiversity Patterns and Processes on the Mid-Atlantic
Ridge. In: McIntyre, A. (Ed.), Life in the World’s Oceans: Diversity, Distribution, and
Abundance, Wiley-Blackwell, Oxford, UK. Pp. 103-121.
Webb, T.J., Berghe, E.V., O’Dor, R., 2010. Biodiversity’s Big Wet Secret: The Global
Distribution of Marine Biological Records Reveals Chronic Under-Exploration of the
Deep Pelagic Ocean. PLoS ONE 5(8), 1-6.
36
Wenneck, T. de Lange, Falkenhaug, T., Bergstad, O.A., 2008. Strategies, methods, and
technologies adopted on the RV G.O. Sars MAR-ECO expedition to the Mid-Atlantic
Ridge in 2004. Deep-Sea Research II 55, 6-28.
Wilson, D.W., 1992. Interactions of ocean currents and diel migrators at a seamount in the
central north Pacific Ocean. PhD Dissertation, University of Hawaii, Honolulu, HI.
Youngbluth, M., Sørnes, T., Hosia, A., Stemmann, L., 2008. Vertical distribution and relative
abundance of gelatinous zooplankton, in situ observations near the Mid-Atlantic Ridge.
Deep-Sea Research II 55, 119-125.
Zuur, A.F., Ieno, E.N., Elphick, C.S., 2010. A protocol for data exploration to avoid common
statistical problems. Methods in Ecology & Evolution 1(1), 3-14.
37
Table 1. Fit statistics associated with the seven models fitted to the Bathylagus euryops catch-per-uniteffort data (individuals / 103 • m3collected at the northern Mid-Atlantic Ridge during the 2004 RV
G.O. Sars expedition: model number, explanatory variables included in the model, -2 log( L̂ ), number
of parameters, Akaike's Information Criterion for small sample size (AICc), and ΔAICc. Depth
corresponds to the five depth zones sampled including 0-200 m, 200-750 m, 750-1500 m, 1500-2300
m, and >2300 m. Ridge corresponds to the four ridge sections sampled including Reykjanes Ridge,
Charlie-Gibbs Fracture Zone, Faraday Seamount Zone, and Azorean Zone. Prey abundance (was
derived at each depth zone/ridge section combination utilizing ‘Krill’ trawl catch data for amphipods,
decapod shrimp, euphausiids, and cnidarians.
Model
no.
1
2
3
4
5
6
7
Explanatory variables
-2 log( L̂ )
Depth
Ridge
Depth, Ridge
Depth, Ridge, Depth*Ridge
Depth, Ridge, Prey
Depth, Ridge, Prey, Depth*Prey
Depth, Ridge, Prey, Depth*Prey, Depth*Ridge
38
308.2
308.8
272.1
249.0
270.0
265.6
238.1
No. of
Parameters
6
5
9
19
10
14
24
AICc
321.2
319.5
292.7
301.5
293.2
300.6
312.4
∆AICc
28.5
26.8
0.0
8.8
0.5
7.9
19.7
Table 2. Parameter estimates, estimated standard errors (SE), t-values associated with the test of the null
hypothesis that βi is equal to zero, and corresponding p-values obtained from model M3 fitted to
Bathylagus euryops catch-per-unit-effort data (individuals·10-3·m-3) collected at the northern Mid-Atlantic
Ridge during the 2004 RV G.O. Sars expedition. The parameter β0 is the reference level of the depth and
ridge section variables and defined to be depth zone 1 (epipelagic) within the Azorean Zone, the
parameters βd2 through βd5 are the estimated effects on CPUE for depth zones 2-5 (200-750 m, 750-1500
m, 1500-2300 m, and >2300 m, respectively) relative to the depth zone 1, and the parameters βCGFZ, βFSZ,
and βRR are the estimated effects on CPUE for ridge sections Charlie-Gibbs Fracture Zone, Faraday
Seamount Zone, and Reykjanes Ridge relative to the Azorean Zone. The estimated residual error, σˆ 2 ,
was 4.12.
Parameter
β0
βd2
βd3
βd4
βd5
βCGFZ
βFSZ
βRR
Estimate
-11.03
2.91
4.11
4.07
1.04
3.95
3.19
3.60
SE
0.67
0.72
0.76
0.76
1.03
0.67
0.85
0.69
t-stat
-16.41
4.05
5.43
5.38
1.01
5.92
3.74
5.22
39
p-value
< 0.001
< 0.001
< 0.001
< 0.001
0.31
< 0.001
< 0.001
< 0.001
Table 3. Parameter estimates, estimated standard errors (SE), t-values associated with the test of the null
hypothesis that βi is equal to zero, and corresponding p-values obtained from diel vertical migration
model fitted to Bathylagus euryops catch-per-unit-effort data (individuals·10-3·m-3) collected at the
Charlie-Gibbs Fracture Zone during the 2009 FSV Henry B. Bigelow expedition. The parameter β0 is the
reference level of the time of day and depth variables and defined to be depth zone 1 (epipelagic) during
the daytime, the parameters βd2 through βd5 are the estimated effects on loge(CPUE) for depth zones 2-5
(200-750 m, 750-1500 m, 1500-2300 m, and >2300 m, respectively) relative to the depth zone 1, and the
parameter βnight is the estimated effect on loge(CPUE) at night relative to the daytime.
Parameter
Estimate
SE
t-stat
p-value
β0
-5.12
1.53
-3.35
< 0.05
βnight
-0.69
1.17
-0.59
0.56
βd2
2.61
1.67
1.56
0.13
βd3
3.73
1.74
2.14
< 0.05
βd4
2.42
2.08
1.16
0.25
βd5
-0.56
2.08
-0.27
0.79
40
Table 4. Parameter estimates, estimated standard errors (SE), t-values associated with the test of the null
hypothesis that βi is equal to zero, and corresponding p-values obtained from ontogenetic vertical
migration model fitted to Bathylagus euryops standard lengths (mm) collected at the Charlie-Gibbs
Fracture Zone during the 2009 FSV Henry B. Bigelow expedition. The parameter β0 is the reference level
of the depth variable and defined to be depth zone 1 (epipelagic), the parameters βd2 through βd5 are the
estimated effects on standard lengths for depth zones 2-5 (200-750 m, 750-1500 m, 1500-2300 m, and
>2300 m, respectively) relative to the depth zone 1.
Parameter
Estimate
SE
t-stat
p-value
β0
βd2
βd3
βd4
βd5
58.87
35.90
52.93
96.48
96.86
25.21
26.09
25.56
27.96
33.34
2.34
1.38
2.07
3.45
2.91
< 0.05
0.17
< 0.05
< 0.05
< 0.05
41
Figure 1. Trawl sampling stations for Leg 1 of 2004 R/V G.O. Sars MAR-ECO expedition to the MidAtlantic Ridge, from Iceland to the Azores (adapted from Sutton et al., 2008).
42
Figure 2. Cruise path and trawl sampling locations during the 2009 FSV Henry B.
Bigelow MAR-ECO expedition to the Charlie-Gibbs Fracture Zone. Box at right shows
the northern Mid-Atlantic Ridge, from Iceland to the Azores, with Charlie-Gibbs Fracture
Zone offset at the center.
43
3
3
Figure 3. Predicted catch (individuals / 10 • m ) of Bathylagus euryops per depth zone at each ridge
section sampled along the MAR during the 2004 RV G.O. Sars expedition. Predicted catch based on
GLM regression containing depth zone and ridge section as explanatory variables. Error bars represent
standard error of the predicted CPUE values at each depth zone relative to the four ridge sections. Note:
scale of the x axis at the Azorean Zone is an order of magnitude smaller than the other ridge sections.
44
Figure 4. Predicted mean standard length of Bathylagus euryops relative to the depth zones sampled at
the Charlie-Gibbs Fracture Zone. Error bars represent standard error of the predicted mean standard
length values relative to depth zone. Results based on 46 discrete-depth ‘macrozooplankton’ trawl
samples from 2009 FSV Henry B. Bigelow expedition.
45
CHAPTER 2
FEEDING ECOLOGY OF BATHYLAGUS EURYOPS (TELEOSTEI: MICROSTOMATIDAE)
ALONG THE NORTHERN MID-ATLANTIC RIDGE
46
ABSTRACT
The northern Mid-Atlantic Ridge, from Iceland to the Azores (MAR), is a massive
ecosystem, yet despite its size few studies have described the feeding ecology of pelagic fishes
along the ridge. Recent efforts from MAR-ECO, a Census of Marine Life field project, aimed to
address this lack of knowledge. Among the midwater fishes sampled during the MAR-ECO
campaign, Bathylagus euryops (Argentiniformes: Microstomatidae) was the biomass-dominant
species and ranked third in total abundance. In this paper, we describe the feeding ecology of B.
euryops as a function of life history and spatial parameters. Furthermore, daily ration estimates
were generated based on different temperature and geographic location combinations.
Multivariate analyses, including a cluster analysis and a canonical correspondence analysis,
revealed that fish size and geographic location significantly influenced the diet of B. euryops.
Furthermore, daily rations were consistently estimated to be less than 1% of the average wet
weight of B. euryops along the MAR. The general trend observed was a southward increase in
daily ration estimates along the MAR. Due to its high abundance and biomass observed along
the MAR, B. euryops appears to be an important species in the oceanic food web of the North
Atlantic Ocean.
47
I. Introduction
In the deep pelagic realm, little is known about how organisms acquire food necessary to
sustain their metabolic needs. Knowledge about trophic interactions is the key to understanding
how energy is transferred from the productive epipelagic zone to the meso- and bathypelagic
zones. Because primary production is concentrated in the epipelagic zone, except in
chemosynthetic systems, all life in the ocean is supported by the downward transfer of organic
matter from the euphotic zone (Angel 1997). Because most productivity occurs in surface
waters, sinking particle flux and biomass is thought to decrease exponentially with increasing
depth from the surface; (Angel & Baker 1982; Reinthaler et al. 2006; Arístegui et al. 2009;
Robinson et al. 2010). Thus, deep-sea ecosystems are generally perceived as food-limited
environments.
Seamount and mid-ocean ridge ecosystems represent a deviation from the classic
paradigm of deep-sea habitats. These physical structures act as obstacles to water circulation and
can induce local currents. Abrupt topographic features can also enhance upwelling and may
bring nutrients up from the deeper ocean, potentially increasing primary production and
supporting a diversity of life (Rogers 1994). To support higher trophic levels, the injection of
nutrients from upwelling must penetrate the photic later and remain long enough to affect
zooplankton production (weeks to months), but evidence of this is limited (Genin & Dower
2007). Regardless, the importance of seamounts and mid-ocean ridges in supporting high fish
biomass is being increasingly recognized. The apparent ecological importance of seamounts and
mid-ocean ridges to fishes and other taxa, in part, inspired the initiation of the ‘Patterns and
Processes of the Ecosystem of the Northern mid-Atlantic’ (MAR-ECO). MAR-ECO is a Census
of Marine Life field project designed to describe and understand patterns of distribution,
48
abundance, and trophic relationships of organisms inhabiting the northern Mid-Atlantic Ridge,
from Iceland to the Azores (MAR hereafter). Recent observations from MAR-ECO suggested a
water column maximum in fish biomass between 1500 – 2300 m (Sutton et al. 2008). This
finding stands in contrast to the traditional view that biomass declines exponentially with
increasing depth from the surface.
Among the fishes sampled along the MAR, Bathylagus euryops (Argentiniformes:
Microstomatidae) was the biomass dominant midwater species and ranked third in total
abundance (Sutton et al. 2008). The trophic dynamics of B. euryops, like most deep-sea fishes,
are poorly understood. Several dietary analyses have been conducted on this species (Mauchline
& Gordon 1983; Klimpel et al. 2006; Busch et al. 2008), however, a majority of the prey items
were deemed unidentifiable in each study. A clearer resolution of the diet of this abundant
species is desired to facilitate an understanding of food web interactions in the MAR.
Bathylagus euryops is distributed throughout the North Atlantic Ocean, from Greenland to
Bermuda, predominantly in the meso- to bathypelagic depths. However, recent observations
have demonstrated that B. euryops may exhibit an ontogenetic vertical migration which likely
explains why smaller fish have been found in the surface waters (Sweetman et al., in prep.).
Although there are no current fisheries for B. euryops, it is an important prey type for higher
trophic levels, including commercially important species such as orange roughy (Hoplostethus
atlanticus), alfonsino (Beryx splendens), and Greenland halibut (Reinharditus hippoglossoides)
(Mauchline & Gordon 1984; Chambers & Dick 2007; Horn et al. 2010).
In general, midwater fishes serve as an important link between zooplankton and higher
trophic level predators such as squids, piscivorous fishes, seabirds, and marine mammals
(Porteiro & Sutton 2007). However, the ecosystem processes that govern the trophic dynamics
49
of fishes along seamounts and mid-ocean ridges are not fully understood. Several mechanistic
hypotheses have been postulated to explain how large aggregations of midwater fishes can be
supported in what is typically considered a food-limited environment. Initially proposed by
Isaacs & Schwartzlose (1965) and later corroborated by a series of studies over several shallow
and intermediate seamounts (Genin et al. 1988, 1994; Fock et al. 2002), the ‘topographic
blockage hypothesis’ describes how the trapping of vertically migrating zooplankton can provide
up to 40 times more food to seamount communities than export (via sinking particles) of
overlying primary production. Numerous species of oceanic zooplankton migrate daily between
the mesopelagic and epipelagic zones. Some vertically migrating zooplankton can reach 800 m
depth (Wiebe et al. 1979), but in the open ocean most reside in the upper 300-600 meters during
the day and migrate to surface waters at night (Steinberg et al. 2008). The summits of many
seamounts and mid-ocean ridges are well within the daily vertical migration range of
zooplankton. If the summit is shallower than preferred daytime depths of zooplankton, vertical
migration would become inhibited, resulting in trapping and increased vulnerability to predators
(Genin & Dower 2007). The topographic trapping of zooplankton would be most effective at
summits in the shallow to intermediate depth range and can greatly improve food availability to
zooplanktivorous fishes (Nonacs et al. 1994; Kiflawi & Genin 1997). Although this mechanism
would most benefit site-attached demersal fishes, it may also have key implications for pelagic
zooplanktivorous fishes like Bathylagus euryops. Wilson (1992) demonstrated that Maurolocus
muelleri and other micronekton, despite strong currents, were able to maintain their position
above the Southeast Hancock Seamount. These strong currents, in theory, should have displaced
the organisms to the surrounding deep waters. The ability for the fish and micronekton
aggregrations to maintain position above the Southeast Hancock Seamount, combined with the
50
observed year-long increase in body size of an apparent cohort of M. muelleri (Boehlert et al.
1992), indicated that these aggregations were resident (Genin 2004). Considering the itinerant
behavior characteristic of pelagic fishes, an increase in energy expenditure relative to demersal
fishes would likely be required for pelagic species to maintain proximity to structures such as
seamounts. For the topographic blockage mechanism to hold true for pelagic fishes, energy lost
in maintaining position against strong currents would need to be more than compensated by an
enhanced trophic subsidy.
Determining biological and environmental patterns in the feeding ecology of deep-sea
fishes remains problematic due to technological challenges of sampling the environment, and
these challenges are further exacerbated when sampling over abrupt topography. Most studies to
date have focused on descriptions of diet, feeding periodicity patterns, and prey selectivity of
deep-pelagic fishes. Fishes that reside in the mesopelagic and bathypelagic zones are typically
classified into three trophic guilds: zooplanktivores, micronektivores, and generalists (Gartner
1997). However, changes in diet and food web interactions occur as fishes continue to grow
(Wootton 1998), and these ontogenetic changes are critical to understanding the ecology of
fishes (Werner and Gilliam 1984) and when possible should be considered in dietary analyses.
Quantitative studies on food consumption by fishes represent a critical step towards
understanding trophic dynamics of pelagic marine communities. Although extensive literature
on the diet composition of midwater fishes is available (Collard 1970; Mauchline and Gordon
1985, 1986), the majority of analyses are non-quantitative in terms of trophic exchanges. Gastric
evacuation and daily ration models have been widely applied to coastal fish communities to
estimate consumption (Elliott and Persson 1978; Durbin et al. 1983), and the application of these
51
models to deep-sea species are needed to advance from a qualitative to a quantitative assessment
of trophic relationships in the deep sea.
The high biomass and abundance of Bathylagus euryops observed over the MAR
suggests their significant ecological importance in this ecosystem. In this paper, we describe the
diet composition of B. euryops over the MAR as the percent composition by weight (mg) (% W)
and the percent frequency of occurrence (% F) of each prey group. Variation in the diet of B.
euryops is also investigated relative to ontogeny and environmental parameters, such as
geographic location and depth. Additionally, we provide estimates of the daily ration of B.
euryops to further elucidate its role in the food web as the biomass dominant species along the
MAR.
II. Materials and methods
2.1. Field collections
Specimens were collected during Leg 1 of the R/V G.O. Sars expedition, 5 June – 3 July,
2004. Sampling occurred at 18 predefined stations along the MAR with coverage over four
different ridge sections (Fig. 1): Reykjanes Ridge (RR), Charlie-Gibbs Fracture Zone (CGFZ),
Faraday Seamount Zone (FSZ), and Azorean Zone (AZ). Vertical temperature profiles were
generated from the surface to near-bottom with a CTD at all stations. Further methodological
details for environmental data collection are provided in Søiland et al. (2008).
Samples were collected using a double-warp midwater ‘Krill’ trawl deployed with
standard pelagic trawl doors. The ‘Krill’ trawl had a mouth area of 36 m2 (6 × 6 m2) and a net
with a mesh size of 3 × 3 mm (6 mm stretched). Total trawl length was 45 m and the trawl body
was equipped with a multi-sampler (remotely operated multiple cod ends) consisting of five 30
52
m long cod ends that allowed for five discrete-depth samples per tow. Additionally, the trawl
was equipped with SCANMAR sensors to provide data on net mouth configuration and net
depth. Oblique tows were performed, from 3000 m to the surface, and nets were remotely
opened and closed at discrete-depth intervals. The trawl was towed at an average speed of 2.2
knots and retrieved at a rate of 25 m min-1. Sampling duration at each depth stratum ranged
between 10 – 40 minutes. The volume of water filtered was subsequently calculated for each
cod end based on the door spread distance, speed of the ship during each sampling event, and
duration of each sampling event to provide catch-per-unit-effort (CPUE) measurements.
Catches from each cod end were sorted separately to prevent mixing of specimens. The
deepest net was usually sorted first, while subsequent catches were stored in a cold room to
prevent sample degradation. Bathylagus euryops from each cod end were enumerated and
weighed in bulk (± 0.1 g) on a motion-compensating scale. Specimens were then frozen in lots
at -20 ⁰C prior to storage at the Bergen Museum of Zoology, http://collections.uib.no/vertebrate/,
the permanent repository for all pelagic fishes collected during the 2004 MAR-ECO expedition.
2.2. Identification of stomach contents
A total of 221 Bathylagus euryops specimens were selected for microscopic stomach
content analysis. Bulk lots of B. euryops were selected based on ridge section, depth zone, and
time of day. Specimens were randomly selected within the lots for use in this study. Standard
length, from the most anterior portion of the premaxilla to the origin of the caudal fin, was
recorded for each fish to the nearest 0.1 mm. An incision was made at the isthmus, continuing
along the ventral midline, and ended at the anus to expose the internal organs. The most anterior
portion of the stomach was cut, such that the liver was separated from the stomach. The
53
intestines were then cut as posteriorly as possible and the entire digestive tract was removed
from each specimen. The contents of the stomach and intestine were removed separately. Small
prey items were mixed in a glycerin and methylene blue solution, placed on a microscope slide,
and sealed with Perimount for long term preservation. Larger prey items were placed in vials
and stored in 70% ethanol.
Prey items that were highly masticated and/or digested prevented taxonomic resolution to
species level. Consequently, prey types were classified into broad taxonomic levels in order to
reduce potential biases in prey identification. Stomach and intestine contents were identified
using a Nikon Eclipse 80i compound microscope for small prey items and a Nikon SMZ 1500
stereo microscope for large prey items. Prey encountered in the esophagus and buccal cavity
were not included for identification to avoid possible bias due to post-capture ingestion (i.e., net
feeding). To avoid the possibility of double-counting prey items within a stomach, organisms
were considered separate only if a particular diagnostic feature was observed multiple times. For
prey-biomass determination, length and width measurements were taken for each identified prey
item using the Nikon Imaging Software Elements BR v.3.2. Body volume was estimated
utilizing volumetric equations of shapes similar to that of the prey taxa (Table 1). Body volume
was converted to wet weight under the assumption of specific gravity equivalence to one gram
cm-3.
2.3. General diet description
To summarize the diet composition of Bathylagus euryops along the MAR, dietary
indices of percent composition by weight and percent frequency of occurrence were calculated
for each prey type following Hyslop (1980). Dietary index % Ij, which can be the percent weight
54
or frequency of occurrence, was calculated using a cluster sampling (CS) estimator (Bogstad et
al. 1995; Buckel et al. 1999; Latour et al. 2008) of the form:
n
%I j =
∑M
i =1
n
ik
∑M
i =1
qik
× 100
(1)
ik
such that,
nk
qˆik =
∑w
k =1
n j nk
ijk
,
(2)
∑∑ wijk
j =1 k =1
where n is the number of tows containing B. euryops, Mik is the CPUE of B. euryops collected
during each sampling event, nj is the total number of prey types observed in the stomachs of all
B. euryops examined, and nk is the total number of positive stomachs examined. A sampling
event was defined as a discrete-depth that was sampled at each station, whereas a positive
stomach was defined as a gastro-intestinal tract that contained prey. In equation (2), q̂ik
represents the proportion of weight or occurrence of each prey type during each sampling event.
To ensure that the data are scaled to represent population-level predatory impacts, the diet
composition indices were calculated as a weighted average of q̂ik with the abundance of B.
euryops representing the scaling component. The variance estimate for % I j was given by
n
var(%I j ) =
∑M
1
nM
2
i =1
ik
(qik − I j ) 2
n −1
× 100 2
where M is the average number of B. euryops collected during a sampling event.
55
(3)
2.4. Ontogenetic and spatial changes in diet
Body size has a major influence on an organism’s metabolic requirements and its
potential for resource exploitation. As a result, the ecological roles of different size conspecifics
may be drastically dissimilar. To better understand the effect of fish size on diet, we calculated
the weights of the diets of Bathylagus euryops (using Eq. 1) and grouped fishes into 30 mm size
bins. An agglomerative hierarchical cluster analysis was conducted to identify dietary
similarities among size categories, including empty stomachs, using a distance matrix of fish size
bins and weights of different prey groups. Each size bin of B. euryops begins as a single object,
and objects that are close together are merged into a cluster. “Close” was determined using the
average linkage method, in which distance values between objects and clusters are computed as
the average dissimilarity between clusters. Bray-Curtis dissimilarity distance coefficients were
utilized to measure distances between objects. Ultimately, size groups were sequentially
combined and a dietary interrelatedness of the different size-classes of B. euryops along the
MAR was inferred from the resulting dendrogram. Ontogenetic changes in the diet determined
from the cluster analysis were then used in subsequent analyses.
Canonical correspondence analysis (CCA; ter Braak 1986) was used to explore the
relationship between the diet of Bathylagus euryops and three explanatory variables: ridge
section (RR, CGFZ, FSZ, AZ), depth zone (0-200 m, 200-750 m, 750-1500 m, 1500-2300 m,
>2300 m) and fish size (small, medium, large, as derived from the hierarchical cluster analysis).
CCA ordination extracts the major gradients in a multivariate dataset (i.e., diet) and explains the
gradients based on different combinations of explanatory variables (McGarigal et al. 2000).
Each element of the response matrix was the mean percent weight of a given prey type within a
56
particular ridge section, depth zone, and fish size combination, so that the amount of variability
in B. euryops diet was explained by the canonical axes. Significance of the explanatory variables
was determined using permutation tests (ter Braak, 1986) and a prey group – explanatory factor
biplot was generated to further explore dietary trends. Detailed diet descriptions were then
developed for significant explanatory variables.
2.5. Gastric evacuation rate and daily ration
Gastric evacuation rates (R) and daily rations are typically obtained through laboratory
starvation experiments. Due to the limitations of determining gastric evacuation rates in the lab
for most deep-sea fishes, independent estimates of R were needed in order to calculate a daily
ration. Numerous studies have concluded that the gastric evacuation rate in fishes is greatly
affected by water temperature (Elliott and Persson 1978; Bromley 1994). Evacuation rates tend
to increase with temperature and the relationship between the two usually follows an exponential
curve (Elliott 1991) and can be estimated using the following equation:
R = aebT
(4)
where R is the gastric evacuation rate, a and b are constants, and T is temperature (⁰ C).
Temperature data from the G.O. Sars expedition, combined with a range of constants
derived from the primary literature, were used to determine a range of potential gastric
evacuation rates for Bathylagus euryops along the MAR. Constants a and b were selected from
species that were either phylogenetically related to B. euryops (Superorder Protacanthopterygii,
Order Salmoniformes) or species that inhabit similar environments (Order Stomiiformes). The
species for which constants were used within the salmonids included Salmo trutta, Salvelinus
57
alpinus, and Onchorynchus nerka (Elliot 1972; Doble and Eggers 1978; Amundsen and
Klemetsen 1988), which all predominantly live in shallow water ecosystems. The species for
which constants were used from the Order Stomiiformes included Valenciennellus tripunctatus,
Vinciguerria nimbaria, and Danaphos oculatus (Pakhomov et al. 1996). Gastric evacuation rates
were determined at each ridge section and incorporated temperatures in which B. euryops were
sampled utilizing Eq. 4, and a range of daily rations were subsequently estimated following
Elliott and Persson’s (1978) model:
Daily rationi = 24 Si R
(5)
where Si is the mean amount of food in the stomach of B. euryops at ridge section i, 24
represents the number of hours in a day, and R is the gastric evacuation rate (Tyrell et al. 2007;
Garrison et al. 2010). Due to the underlying assumptions of applying a and b constants derived
from different species, a sensitivity analysis was performed to understand the influence each had
on the overall magnitude of the daily ration of B. euryops. Ultimately, the highest and lowest R
values were used to estimate a high and low range of daily ration estimates at each of the ridge
section and temperature combinations. A percent difference between the high and low daily
ration estimates was then calculated to quantify the influence that the different R values used in
this study had on the daily rations.
III. RESULTS
3.1. General diet description
During the 2004 R/V G.O. Sars expedition, Bathylagus euryops were collected during 46
sampling events, and at least one B. euryops stomach from 35 sampling events contained prey.
58
Prey were observed in 112 (50.7%) of the 221 stomachs collected. Twenty-nine prey types were
observed in the diet of B. euryops, 20 of which were identifiable to family level and 19
identifiable to genus level.
The percent-wet weight dietary index indicated that copepods were the primary prey of
Bathylagus euryops along the MAR, accounting for 37.8% of the diet (Fig. 2A). Decapod
shrimp ranked second in dietary importance by weight and accounted for 20.5% of the diet.
Gelatinous zooplankton (15.2%) and amphipods, mainly of the Suborder Hyperiidea (13.8%),
were of tertiary importance. Ostracods were the only other prey group that contributed more
than 5% of the diet by weight (8.7%). Remaining measurable and identifiable prey types,
including chaetognaths and echinoderms, collectively represented less than 2% of the diet.
The percent-frequency-of-occurrence index also revealed that copepods were the major
prey of Bathylagus euryops, occurring in 33.6% of positive stomachs (Fig. 2B). Ostracods were
also important in the diet of B. euryops, occurring in 20.9% of positive stomachs. Amphipods
(8.4%) and decapod shrimp (7.8%) ranked third and fourth in dietary importance based on the
frequency of occurrence index. Chaetognaths and gelatinous zooplankton were nearly equally
important at 5.9% and 5.8%, respectively. Other identifiable prey types including radiolarians,
forams, gastropods, echinoderms, and cephalopods were found in 4.7% of positive stomachs
combined. Unidentifiable prey items (i.e., crustacea fragments and unidentifiable material) were
prevalent, occurring in 12.5% of fish containing prey.
3.2. Ontogenetic and spatial changes in diet
Specimens of Bathylagus euryops used in this study ranged in size from 40.2 to 215.0
mm SL (Fig. 3). Observations from the largest size bin (185 – 215 mm SL) were excluded from
59
the hierarchical cluster analysis due to low sample size in an effort to minimize the influence of
outliers (McGarigal et al. 2000). A scree plot indicated three clusters (Fig. 4A) corresponding to
three size-categories: < 95 mm (small), 95 – 155 mm (medium) and > 155 mm SL (large) (Fig.
4B). Canonical correspondence analysis indicated that dietary changes by fish size (F = 1.97, p
= 0.05) and ridge section (F = 1.83, p = 0.10) were significant at α = 0.10, whereas depth zone (F
= 0.64, p = 0.88) was not significant. Collectively, the explanatory variables accounted for
43.4% of the variability in the diet. The first and second canonical axes accounted for 63.1% and
31.8% of the explainable variation, respectively. Fish size more closely corresponded to the first
canonical axis than the second, and represented the greatest component of diet variation.
Conversely, ridge section and depth zone were more closely correlated to the second axis (Fig.
5).
Ostracods represented the dominant prey group in the small size class of Bathylagus
euryops (< 95 mm SL), accounting for 49.1% of the diet by weight (Fig. 6A). Hyperiid
amphipods ranked second in dietary importance at 28.6% of the diet by weight. Gelatinous
zooplankton accounted for 17.5% of the diet, and copepods (4.8%) from the Order Calanoida
were the only other prey group observed in the diet of small B. euryops. The diet composition of
the medium size class of B. euryops (95 – 155 mm SL) was dominated by copepods, accounting
for 32.6% of the diet by weight (Fig. 6B). Gelatinous zooplankton ranked second in dietary
importance (24.5%), and decapod shrimp accounted for 14.7% of medium sized fish diet by
weight, followed by amphipods (12.0%) and ostracods (7.3%). Diet of the large size class (>155
mm SL) was dominated by decapod shrimp and copepods, which represented more than 80% of
diet by weight (47.5% and 35.5%, respectively, Fig 6C). Amphipods (13.1%), ostracods (2.2%),
and gelatinous zooplankton (1.7%) contributed to the remaining diet weight profile.
60
Spatial changes in the diet of Bathylagus euryops was detected between the northern
ridge sections (RR and CGFZ) and the southern ridge sections (FSZ and AZ) (Fig. 9). In the
northern ridge sections, copepods (43.0%) and ostracods (15.6%) were more prevalent in the
diets when compared to the southern ridge sections (6.0% and 0.2%, respectively). Among the
copepods found in the diet of B. euryops along the northern ridge sections, Calanus finmarchicus
and Microcalanus spp. were the most prevalent taxa. The amount of amphipods and decapod
shrimps in the diets of B. euryops remained relatively consistent regardless of ridge section.
Gelatinous zooplankton were the only prey group more prevalent in the diet of B. euryops from
southern ridge sections (36.4%) relative to the northern ridge sections (5.8%). Chaetognaths and
echinoderms were observed in approximately 2% of the diet by weight in the northern ridge
sections, but were not present in the diets of fish sampled in the southern ridge sections.
3.3. Gastric evacuation and daily ration estimation
Bathylagus euryops were sampled in water temperatures ranging from 3 – 9 ⁰ C along the
MAR, except in the Azorean Zone where fish were only collected in temperatures between 3 – 5
⁰ C. Overall, daily rations were highest at 9 ⁰ C and lowest at 3 ⁰ C (Table 2). The average
weight of food in the stomach of B. euryops over a 24 hour period was highest at the
southernmost ridge section, the Azorean Zone (98.13 mg wet weight). To the north, the average
weight of prey in stomachs decreased to 83.64 mg wet weight at the Faraday Seamount Zone to
30.02 mg wet weight at the Charlie-Gibbs Fracture Zone. The average weight of prey in the
stomachs of B. euryops was lowest at the Reykjanes Ridge (12.54 mg wet weight). Regardless
of ridge section, daily ration estimates never exceeded 1% of the average wet weight of B.
euryops per day. When ridge section was considered, daily rations were highest at the
61
southernmost ridge section (Azorean Zone) and lowest at the northernmost ridge section
(Reykjanes Ridge). Results from a sensitivity analysis revealed that the different constants used
in Eq. 4 had minimal influence on the magnitude of the daily ration estimates. Overall, high and
low daily ration estimates differed at the most by only 0.63%.
IV. DISCUSSION
4.1. General diet description
Bathylagus euryops feed on a diverse array of prey in the North Atlantic Ocean,
evidenced by the different taxa encountered in the diet in this study. Fishes from the Family
Microstomatidae are typically classified as zooplanktivores that specialize on gelatinous
zooplankton (Hopkins et al., 1996; Gartner Jr. et al., 1997; Arai, 2005). However, this study
found that gelatinous zooplankton represented 15.2% of the diet by (wet) weight and occurred in
only 5.9% of stomachs. While the rapid digestion of gelatinous zooplankton in fish stomachs
undoubtedly lead to an underestimation of their overall contribution to diet, methodical
discrepancies between study calculations also represent a plausible explanation for the observed
differences.
In general, it is difficult to compare studies of diet composition of the same species due to
differences in survey design, dietary indices reported, and methods used to calculate these
indices. Overall, crustaceans dominated the diet of Bathylagus euryops, with copepods ranking
first in magnitude across the dietary indices generated in this study. Copepods are the
numerically dominant members of marine plankton at all depths in the ocean (Herring 2002) and
their importance in the diet of deep-sea fishes has been widely documented (Hopkins and Baird
1977). Mauchline and Gordon (1983) characterized the diet of B. euryops by generating a
62
frequency of occurrence dietary index and found that calanoid copepods were the main prey
group observed in the stomach contents with identifiable food, and our results agree with the
findings in this study. Klimpel et al. (2006) and Busch et al. (2008) also investigated the diet of
B. euryops and reported percent-composition-by-weight dietary indices for unidentified
crustaceans that ranged from 77.9 – 92.3%. While these studies were mainly focused on the
occurrence of fish parasites in the deep-sea, and consequently provided low taxonomic resolution
of diet, our results corroborate the importance of crustaceans in the diet of B. euryops. In the
present study, unidentifiable prey items were observed in the stomachs where identifiable prey
items were also found, and were likely the same prey types as the latter. However, these prey
items were classified as unidentifiable to provide a conservative diet description.
4.2. SRS vs. CS estimation methods
In the present study we utilized a cluster sampling estimator in contrast to past use of the
simple random sampling (SRS) estimator (e.g. Mauchline and Gordon 1983; Hopkins et al.
1996). By definition, SRS assumes each fish has an equal probability of being sampled out of a
population of N fish (Cochran 1977). Accordingly, each fish sampled is assumed to be an
independent sampling unit. However, fishes and their prey are known to distribute in spatially
variable aggregations (Pitcher and Parrish 1993) as a result of biotic interactions and tolerances
to physical conditions, including but not limited to, resource partitioning, food availability,
reproduction, predation impacts, and temporal changes in the physical environment. Since fishes
are typically not distributed evenly, the SRS assumption of equal probability of capture is
frequently violated. Furthermore, zooplankton patch dynamics over abrupt topography may
result in correlation among stomach content data within planktivorous fishes sampled at a single
63
location as it is metabolically cheaper to feed on patches of prey than on evenly distributed prey
(Horwood and Cushing 1978). Thus, sampling using active gears may yield randomly collected
clusters of fishes rather than a sample of randomly selected individuals. A cluster sampling (CS)
estimator is appropriate if fishes are determined to be sampled in clusters, where the diets of
those sampled within a gear deployment are more likely to be similar than those sampled
between gear deployments.
Similarity among diets within and between sampling events can be formally considered
by examining the intracluster correlation coefficient (ICC, ρ), which reflects the relationships of
the variance within a cluster to the variance between clusters (Cochran 1977). ICC estimates can
be produced for each prey type in the diet and values will range from 0 to 1. For circumstances
where ρ is substantially greater than 0 for any prey types, variance between sampling events is
greater than variance within sampling events and the collections of fish should be viewed as
independent sampling units rather than the individuals within a collection (Bogstad et al., 1995).
In this situation, a CS estimator is the most appropriate estimation method. Conversely, in
circumstances where ρ is near 0, variance within a sampling event is substantially greater than
variance between sampling events and a SRS estimator is most appropriate. The variability of B.
euryops stomach contents was greater between samples than within samples as evidenced by
ICC values of 0.49 for amphipods, 0.48 for copepods, 0.59 for gelatinous zooplankton, 0.51 for
decapods, and 0.67 for ostracods. Ultimately, these ICC values indicated that the ‘Krill’ trawl
yielded collections of B. euryops that were likely sampled in clusters and that the application of a
CS estimator (Eq. 1) rather than a SRS estimator is the appropriate method to calculate diet
composition. However, it should be noted that the ICC estimate is used to guide estimation
methods and does not solely determine whether a SRS or CS estimator should be used to
64
generate dietary indices. In addition to the ICC estimate, sampling techniques, the ecology of the
species of concern, and the dynamics that govern the distribution of prey must also be
considered.
The manner in which fishes are sampled is an important consideration in dietary
composition analyses as results may be drastically different with an SRS estimator versus a CS
estimator. For example, if a SRS estimator were applied to this study, the results would conform
to that of previous dietary descriptions for Bathylagus euryops, with gelatinous zooplankton
accounting for about 50% of the diet by weight (as opposed to 15% of the diet by weight with a
CS estimator). The application of the CS estimator essentially weights the stomach contents by
the abundance of fish within a sampling event to ensure accurate population level predatory
impacts. Despite its appropriateness for marine fish diet studies, the utilization of this approach
is rarely reported in the literature, and may therefore entrain inaccurate dietary calculations,
descriptions, and comparisons.
4.3. Ontogenetic and spatial changes in diet
Aside from general diet descriptions, ontogenetic and regional changes in feeding
patterns have been used to explain the underlying variability in the diet of deep-sea fishes
(Hopkins and Baird 1977). Cluster analysis revealed three size classes of fishes that grouped
separately based on their diet composition, and a subsequent CCA determined that fish size
significantly influenced the diet of Bathylagus euryops along the MAR. Small B. euryops fed
predominantly on ostracods, while amphipods and gelatinous zooplankton were of secondary
importance. An ontogenetic expansion in the diet occurred around 95 mm SL, as B. euryops
began feeding on a wider variety of prey groups. Copepods represented the dominant
65
component of the diet in the medium size class, whereas ostracods were no longer a major
component of the diet, and larger prey groups, such as decapod shrimp, began to appear in the
diet. As fish increased in size to larger than 155 mm SL, decapod shrimp dominated the diet by
weight, followed by copepods. Gape limitation, where the limits on food size are imposed by the
size of the mouth opening, likely influenced some of the observed ontogenetic changes in diet.
Some of the smallest prey items, such as ostracods, were most abundant in the diet of smaller
fish. Conversely, decapod shrimp, which were among the largest prey groups, were observed
primarily in the diets of larger B. euryops, which were presumably more capable of capture and
ingestion of the larger prey. An increase in gape limitation would likely allow for a wider
variety of prey groups to be consumed and this was observed between the small size class (35 –
95 mmSL) relative to the medium size class (95 – 155 mm SL). However, a decrease in the
number of prey groups was observed between the medium size class and the large size class (>
155 mm SL) and this was likely an artifact of the small sample size analyzed in large size class
(n = 35).
A central assumption in deep-ocean environments is that opportunistic generalist feeding
strategies should be favored due to relatively sparse food availability, and that this pattern should
be more apparent with increasing depth (Ebeling and Calliet 1974). Results from this study
indicate that ontogenetic changes in the feeding of Bathylagus euryops may contribute to the
observed dietary variation with size. An ontogenetic vertical migration was postulated by
Mauchline and Gordon (1983) and recently observed in B. euryops along the MAR, where
smaller fish were found in surface waters and larger fish at greater depths (Sweetman et al. in
prep.). Dietary differences based on size may be a function of the vertical distribution of
different size classes of B. euryops relative to the vertical distribution of the prey field. For
66
instance, Calanus finmarchicus is the most abundant copepod in the North Atlantic Ocean and
may constitute ~80% of total copepod biomass in the North Atlantic Ocean (Gislason et al.
2008). During the summer months, C. finmarchicus resides in the mesopelagic zone as it
matures in preparation for reproduction during the spring (Sameoto & Herman 1990). Copepods
were a dominant prey group in both the medium and large size classes of B. euryops, and of the
copepods observed in the diet of these size classes, C. finmarchicus was the most abundant. The
depth distribution of C. finmarchicus during the summer likely explains the subsequent increase
in copepod dietary importance with increasing fish size.
Another significant trend observed in the CCA was that most prey groups were more
prevalent in the diet of Bathylagus euryops in the northern ridge sections when compared to the
southern ridge sections, with gelatinous zooplankton representing the only major deviation from
this pattern. The distributions of zooplankton in the open ocean are strongly regulated by, and
spatially correlated with, physical oceanographic processes (Mackas et al. 1985) and the spatial
changes observed in the diet of B. euryops relative to geographic location (e.g., ridge section)
were likely reflective of the hydrographic patterns along the MAR at the time of sampling.
Along the MAR, the Sub-Polar Front (SPF) is a major hydrographic feature and recent findings
from MAR-ECO have shown it to be a biogeographic barrier for several zooplankton taxa
(Falkenhaug et al. 2007). The SPF is a semi-permanent feature with a variable position south of
the CGFZ. The SPF represents a boundary between the cool, lower salinity waters of the north
and the southern warm and saline waters (Rossby 1999). Copepods dominated the diet of B.
euryops in the northern ridge sections. Specifically, Calanus finmarchicus and Microcalanus
spp. were the most frequently observed species in the stomachs of B. euryops from the northern
ridge sections. These copepod taxa were also the most abundant calanoid copepods in waters
67
north of the SPF (Gaard et al. 2008). South of the SPF, abundances of C. finmarchicus and
Microcalanus spp. markedly decline, and a corresponding decrease of copepod dietary
importance was observed in B. euryops. Qualitatively, environmental availability of these
copepod species appears to be proportionally reflected in B. euryops diet.
The other major spatial trend observed in the diet of Bathylagus euryops was the increase
in importance of gelatinous zooplankton in the southern ridge sections. The importance of
gelatinous zooplankton as both predators and prey in the open ocean is becoming increasingly
recognized (Purcell and Arai 2001; Purcell 2005; Link and Ford 2006). However, due to the
fragile tissue structure of these organisms, little information is available on large-scale spatial
patterns of abundance, limiting quantitative inclusion of gelatinous zooplankton in dietary
selectivity analyses. During the MAR-ECO campaign, several different sampling methodologies
were used to elucidate distribution patterns of gelatinous zooplankton, including net sampling,
underwater video profiling, stimulated bioluminescence, and remotely operated vehicles.
Gelatinous zooplankton species richness was highest south of the SPF, where higher abundances
of lobate ctenophores, siphonophores, thaliaceans, and Aglantha spp. were observed compared to
regions north of the SPF (Hosia et al. 2008; Stemmann 2008; Youngbluth 2008). Conversely,
medusae were more abundant along the northern ridge sections. Despite these observations, the
importance of specific taxa of gelatinous zooplankton in the diet of B. euryops remains difficult
to determine due to the highly digested state of the ingested material that prohibited further
taxonomic resolution. The digestion of gelatinous zooplankton in fish stomachs may be nearly
complete in one hour (Arai et al. 2003) and given that most midwater trawls take multiple hours
from deployment to retrieval, additional methods, such as molecular identification, are needed to
68
further understand the importance of gelatinous zooplankton taxa in the diet of B. euryops and
other pelagic fishes along the MAR.
4.4. Gastric evacuation and daily ration estimates
Fish gastric evacuation rates and daily rations are often verified through laboratory
feeding experiments. This application, however, remains problematic in deep-sea research due
to difficulties in keeping deep-sea fishes alive in containment. When calculating R, it is critical
that the constants used reflect species with similar metabolic characteristics that are appropriate
for comparison when the data for the target species are unavailable. In an effort to provide
conservative daily ration estimates for Bathylagus euryops, a and b constants in Eq. 4 were
derived from the literature for phylogenetically-related species and those inhabiting similar
habitats (e.g. deep-sea fishes). Although the a and b constants used in this study were from
phylogenetically and ecologically different species, they did not have a major impact on the
overall daily ration estimates. We determined that daily ration of B. euryops increased
southward along the MAR and never exceeded 1% of the average wet weight per day. Only two
studies have estimated the daily ration for non-migrating fishes within the Family
Microstomatidae (Childress et al. 1980; Gorelova 1985). Although both studies applied different
methods to calculate a daily ration, each determined the daily ration to be very low, not
exceeding 1% of the body weight per day, and our results support these findings. The southward
increase in B. euryops daily ration may be explained by feeding responses correlated with water
temperature. Water temperatures were coldest at the RR and warmest at the AZ, and in the
absence of limiting factors, a temperature increase will increase the metabolic rate of fishes
(Hochachka and Somero 1984). While temperature likely explains much of the variability
69
observed in the daily ration estimates for B. euryops, other factors including social behavior,
food availability, resource partitioning, predation impacts, and responses to environmental
variables other than temperature likely play a role and should be examined further to better
elucidate the feeding capacity of this species.
V. CONCLUSION
In conclusion, the combination of descriptive and quantitative analyses of food web
interactions provides valuable insights into the structure of deep-sea ecosystems along the MAR.
The incorporation of ontogenetic and spatial changes in the dietary analysis of Bathylagus
euryops elucidated some of the ecosystem processes that govern the trophic dynamics of this
species. These findings will ideally aid in the management of commercially important fishes that
ultimately rely on B. euryops as prey. Furthermore, the estimation of a daily ration for B.
euryops represents the first such estimate for the species and our results suggest that this species
is an important consumer of oceanic zooplankton and macroplankton. Due to their high
abundances and biomass observed along the MAR, B. euryops represents an important pathway
in the transfer of carbon from the epipelagic zone to the deep ocean in the mid-oceanic North
Atlantic. However, the patterns described in the present study are based on sampling that took
place in the summer, after the peak in primary production from the spring bloom had already
occurred. The primary remaining gap in the knowledge of the feeding ecology of B. euryops
along the MAR is the influence of temporal patterns (diel, intra-annual, and interannual) and the
influence of hydrography and water masses.
70
ACKNOWLEDGEMENTS
The authors acknowledge the efforts of the MAR-ECO steering committee, particularly
Odd Aksel Bergstad and Tone Falkenhaug, for providing the research platforms which made this
study possible. The crew and field personnel of the RV G.O. Sars deserve thanks for their
contributions in the field. E. Hilton, D. Steinberg, K. Parsons, and J. Kidwell provided helpful
comments on earlier versions of this manuscript. Funding was provided by the Sloan
Foundation/Census of Marine Life and from the NSF Ocean Sciences Division – Biological
Oceanography Program (OCE 0623551) and the VIMS Office of Academic Studies. MAR-ECO
is a Census of Marine Life field project.
71
REFERENCES
Amundsen PA, Klemetsen A (1988) Diet, gastric evacuation rates and food consumption in a
stunted population of Arctic charr, Salvelinus aplinus L., in Takvatn, northern Norway. J
Fish Biol 33:697-709
Angel MV (1985) Vertical migrations in the oceanic realm: possible causes and probably effects.
Contrib Mar Sci 27:45-70
Angel MV (1997) What is the deep sea? In: Randall DJ, Farrell AP (eds) Deep-Sea Fishes.
Academic Press, San Diego
Angel MV, Baker A (1982) Vertical distribution of the standing crop of plankton and
micronekton at three stations in the Northeast Atlantic. Biol Oceanogr 2:1-30
Arai MD, Welch DW, Dunsmuir L, Jacobs MC, Ladouceur AR (2003) Digestion of pelagic
Ctenophora and Cnidaria by fish. Can J Fish Aquat Sci 60:825-829
Arai MD (2005) Predation on pelagic coelenterates: a review. J Mar Biol Assoc UK 85:523-536
Arístegui J, Gasol JM, Duarte CM, Herndl GJ (2009) Microbial oceanography in the dark
ocean’s pelagic realm. Limnol Oceanogr 54(5):1501–1529
Boehlert GW, Watson W, Sun L C (1992) Horizontal and vertical distributions of larval fishes
around an isolated oceanic island in the Tropical Pacific. Deep-Sea Res (1 Oceanogr Res
Pap) 39:439-466
Bogstad B, Pennington M, Volstad JH (1995) Cost-efficient survey designs for estimating food
consumption by fish. Fish Res 23:36-47
Bromley PJ (1994) The role of gastric evacuation experiments in quantifying the feeding rates of
predatory fish. Rev Fish Biol Fish 4:36-66
Buckel JA, Conover DO, Steinberg ND, McKown KA (1999) Impact of age-0 bluefish
(Pomatomus saltatrix) predation on age-0 fishes in the Hudson River estuary: evidence
for density-dependent loss of juvenile striped bass (Morone saxatilis). Can J Fish Aquat
Sci 56:275-287
Busch MW, Klimpel S, Sutton T, Piatkowski U (2008) Parasites of the deep-sea smelt
Bathylagus euryops (Argentiniformes: Microstomatidae) from the Charlie-Gibbs Fracture
Zone (CGFZ). Mar Biol Res 4:313-317
Chambers CA, Dick TA (2007) Using environmental variables to predict the structure of deepsea Arctic fish communities: Implications for food web construction. Arct, Antarct, Alp
Res 39:2-9
72
Childress JJ, Taylor SM, Cailliet GM, Price MH (1980) Patterns of growth energy utilization and
reproduction in some meso- and bathypelagic fishes off Southern California. Mar Biol
76:27-40
Cochran WG (1977) Sampling techniques, 3rd ed. John Wiley & Sons, NY
Collard SB (1970) Forage of some eastern Pacific midwater fishes. Copeia 2:348-354
Doble BD, Eggers DM (1978) Diel feeding chronology, rate of gastric evacuation, daily ration,
and prey selectivity in Lake Washington juvenile sockeye salmon (Onchorynchus nerka).
T Am Fish Soc 107(1):36-45
Durbin EG, Durbin AG, Langton RW, Bowman RE (1983) Stomach contents of silver hake,
Merluccius bilinearis, and Atlantic cod, Gadus morhua, and estimation of their daily
rations. Fish Bull 81(3):437-454
Ebeling AW, Calliet GM (1974) Mouth size and predator strategy of midwater fishes. Deep-Sea
Res 21:1095-1099
Elliott JM (1972) Rates of gastric evacuation in brown trout, Salmo trutta L. Freshwat Biol 2:118
Elliott JM, Persson L, (1978) The estimation of daily rates of food consumption for fish. J Anim
Ecol 47:977-991
Elliott JM (1991) Rates of gastric evacuation in piscivorous brown trout, Salmo trutta. Freshwat
Biol 25:297-305
Falkenhaug T, Gislason A, Gaard E (2007) Vertical distribution and population structure of
copepods along the northern Mid-Atlantic Ridge. ICES ASC, Copenhagen
Fock HO, Matthiessen B, Zodowitz H, von Westerhagen H (2002) Diel and habitat-dependent
resource utilization by deep-sea fishes at the Great Meteor Seamount: niche overlap and
support for sound scattering layer interception hypothesis. Mar Ecol Prog Ser 244:219233
Gaard E, Gislason A, Falkenhaug T, Søiland H, Musaeva E, Vereshchaka A, Vinogradov G
(2008) Horizontal and vertical copepod distribution and abundance on the Mid-Atlantic
Ridge in June 2004. Deep-Sea Res (2 Top Stud Oceanogr) 55:59-71
Gartner JV (1997) Feeding at Depth. In: Randall DJ, Farrell AP (eds) Deep-Sea Fishes.
Academic Press, San Diego
Garrison LP, Link JS, Kilduff DP, Cieri MD, Muffley B, Vaughan DS, Sharov A, Mahmoudi B,
Latour RJ (2010) An expansion of the MSVPA approach for quantifying predator-prey
interactions in exploited fish communities. ICES J Mar Sci 67(5):856-870
73
Genin A, Haury LR, Greenblatt P, (1988) Interactions of migrating zooplankton with shallow
topography: predation by rockfishes and intensification of patchiness. Deep-Sea Res (1
Oceanogr Res Pap) 35:151-175
Genin A, Greene C, Haury LR, Wiebe P, Gal G, Kaartvedt S, Meir E, Feys C, Dawson J (1994)
Zooplankton patch dynamics: daily gap formation over abrupt topography. Deep-Sea Res
(1 Oceanogr Res Pap) 41:941-951
Genin A (2004) Bio-physical coupling in the formation of zooplankton and fish aggregations
over abrupt topographies. J Mar Syst 50:3-20
Genin A, Dower JF (2007) Seamount plankton dynamics. In: Pitcher TJ, Morato T, Hart PJB,
Clark MR, Haggan N, Santos RS (eds) Seamounts: Ecology, Fisheries & Conservation.
Blackwell Publishing, Oxford
Gislason A, Gaard E, Debes H, Falkenhaug T (2008) Abundance, feeding and reproduction of
Calanus finmarchicus in the Irminger Sea and on the northern Mid-Atlantic Ridge in
June. Deep-Sea Res (2 Top Stud Oceanogr) 55:72-82
Gorelova TA (1985) Assessment of the daily ration of meso- and bathypelagic fishes of the
family Bathylagidae. J Ichthyol 25:155-160
Herring P (2002) The marine phyla. In: Crawley MJ, Little C, Southwood TRE, Ulfstrand S
(eds) The Biology of the Deep Ocean. Oxford University Press, New York
Hochachka PW, Somero GN (1984) Biochemical Adaptation: Mechanisms and Process in
Physiological Evolution. Princeton University Press, Princeton
Hopkins TL, Baird RC (1977) Aspects of the feeding ecology of oceanic midwater fishes. In:
Andersen NR, Zahuranec BJ (eds) Oceanic Sound Scattering Prediction. Plenum Press,
New York
Hopkins TL, Sutton TT, Lancraft TM (1996) The trophic structure and predation impact of a low
latitude midwater fish assemblage. Prog Oceanogr 38:205-239
Horn PL, Forman J, Dunn MR (2010) Feeding habits of alfonsino Beryx splendens. J Fish Biol
76:2382-2400
Horwood JW, Cushing DH (1978) Spatial distribution and ecology of pelagic fish. In: Steele JH
(ed) Spatial Pattern in Plankton Communities. Plenum, New York
Hosia A, Stemmann L, Youngbluth M (2008) Distribution of net-collected planktonic cnidarians
along the northern Mid-Atlantic Ridge and their associations with the main water masses.
Deep-Sea Res (2 Top Stud Oceanogr) 55:106-118
74
Hyslop EJ (1980) Stomach contents analysis – a review of methods and their application. J Fish
Biol 17:411-429
Isaacs JD, Schwartzlose RA (1965) Migrant sound scatterers: interaction with the seafloor.
Science 150:1810-1813
Kiflawi M, Genin A (1997) Prey flux manipulation and the feeding rates of reef dwelling
planktivorous fish. Ecology 78:1062-1077
Klimpel S, Palm HV, Busch MW, Kellermanns E, Ruckerta S (2006) Fish parasites in the Arctic
deep-sea: Poor diversity in pelagic fish species vs. heavy parasite load in a demersal fish.
Deep-Sea Research (1 Oceanogr Res Pap) 53:1167-1181
Latour RJ, Gartland J, Bonzek CF, Johnson RA (2008) The trophic dynamics of summer
flounder (Paralichthys dentatus) in Chesapeake Bay. Fish Bull 106:47-57
Link JS, Ford MD (2006) Widespread and persistent increase of Ctenophora in the continental
shelf ecosystem off NE USA. Mar Ecol Prog Ser 320:153-159
Mackas DI, Denman KL, Abbott M.R, (1985) Plankton patchiness: biology in the physical
vernacular. Bull Mar Sci 37:652-674
Mauchline J, Gordon JDM (1983) Diets of clupeoid, stomiatoid, and salmonoid fish of the
Rockall Trough, northeastern Atlantic Ocean. Mar Biol 77:67-78
Mauchline J, Gordon JDM (1984) Occurrence and feeding of berycomorphid and percomorphid
teleost fish in the Rockall Trough. J Cons Int Explor Mer 41:239-247
Mauchline J, Gordon JDM (1985) Trophic diversity in deep-sea fish. J Fish Biol 26:527-535
Mauchline J, Gordon JDM (1986) Foraging strategies of deep-sea fish. Mar Ecol Prog Ser
27:227-238
McGarigal K, Cushman S, Stafford S (2000) Cluster Analysis. Multivariate Statistics for
Wildlife and Ecology Research. Springer, New York
Nonacs P, Smith PE, Bouskila A, Luttbeg, B., 1994. Modeling the behavior of the norther
anchovy, Engraulis mordax, as a schooling predator exploiting patchy prey. Deep-Sea
Res (2 Top Stud Oceanogr) 41:147-169
Pakhomov EA, Perissinotto R, McQuiad CD (1996) Prey composition and daily rations of
myctophid fishes in the Southern Ocean. Mar Ecol Prog Ser 134:1-14
Pitcher TJ, Parrish JK (199) Function of shoaling behavior in teleosts. In: Pitcher TJ (ed)
Behaviour of Teleost Fishes, 2nd ed. Chapman and Hall, London
75
Porteiro FM, Sutton T (2007) Midwater fish assemblages and seamounts. In: Pitcher TJ, Morato
T, Hart PJB, Clark MR, Haggan N, Santos RS (eds) Seamounts: Ecology, Fisheries &
Conservation. Blackwell Publishing, Oxford
Purcell JE, Arai MN (2001) Interactions of pelagic cnidarians and ctenophores with fish: a
review. Hydrobiologia 451:27-44
Purcell JE (2005) Climate effects on formation jellyfish and ctenophore blooms: a review. J Mar
Biol Assoc UK 85:461-476
Reinthaler T, van Aken H, Veth C, Arístegui J, Robinson C, Williams P, Lebaron P, Herndl GJ
(2006) Prokaryotic respiration and production in the meso- and bathypelagic realm of the
eastern and western North Atlantic basin. Limnol Oceanogr 51(3), 1262–1273.
Robinson C, Steinberg DK, Andersen TR, Arístegui J, Carlson CA, Frost JR, Ghiglione JF,
Hernández-Léon S, Jackson GA, Koppelmann R, Quéguiner B, Ragueneau O,
Rassoulzadegan F, Robison BH, Tamburini C, Tanaka T, Wishner KF, Zhang J (2010)
Mesopelagic zone ecology and biogeochemistry – a synthesis. Deep-Sea Res (2 Top Stud
Oceanogr) 57:1504-1518
Rogers A.D (1994) The biology of seamounts. In: Blaxter JHS, Southward AJ (eds) Advances in
Marine Biology, vol 30. Academic Press, San Diego
Rossby T (1999) On gyre interactions. Deep-Sea Res (2 Top Stud. Oceanogr) 46:139-164
Sameoto DD, Herman AW (1990) Life cycle and distribution of Calanus finmarchicus in deep
basins on the Nova Scotia shelf and seasonal changes in Calanus spp. Mar Ecol Prog Ser
66:225-237
Søiland H, Budgell P, Knutsen Ø (2008) The physical oceanographic conditions along the Mid
Atlantic Ridge north of the Azores in June–July 2004. Deep-Sea Res (2 Top Stud
Oceanogr) 55:29-44
Stemmann L, Hosia A, Youngbluth MJ, Søiland H, Picheral M, Gorsky G (2008) Vertical
distribution (0-1000 m) of macrozooplankton, estimated using the Underwater Video
Profiler, in different hydrographic regimes along the northern portion of the Mid-Atlantic
Ridge. Deep-Sea Res (2 Top Stud Oceanogr) 55:94-105
Steinberg DK, Cope JS, Wilson SE, Kobari T (2008). A comparison of mesopelagic
mesozooplankton community structure in the subtropical and subarctic North Pacific
Ocean. Deep-Sea Res (2 Top Stud Oceanogr 55:1615–1635
Sutton TT, Porteiro FM, Heino M, Byrkjedal I, Langhelle G, Anderson CIH, Horne J, Søiland H,
Falkenhaug T, Godø OR, Bergstad OA (2008) Vertical structure, biomass and
topographic association of deep-pelagic fishes in relation to a mid-ocean ridge system.
Deep-Sea Res (2 Top Stud Oceanogr) 55:161-184
76
Sweetman CJ, Sutton T, Vecchione M, Latour RJ (2012) Distribution of Bathylagus euryops
(Teleostei: Bathylagidae) along the northern Mid-Atlantic Ridge. Will submit to DSR part
II or MEPS
ter Braak CJF, (1986) Canonical correspondence analysis: a new eigenvector technique for
multivariate direct gradient analysis. Ecology 67:1167-1179
Tyrell MC, Link JS, Moustahfid H, Smith BE (2007) The dynamic role of pollock (Pollachius
virens) as a predator in the northeast US continental shelf ecosystem: a multi-decadal
perspective. J Northw Atl Fish Sci 38:53-65
Werner EE, Gilliam JF (1984) The ontogenetic niche and species interactions in size-structured
populations. Annu Rev Ecol Evol Syst 15:393-425
Wiebe PH, Nadin LP, Haury LR, Harbison GR, Philbin LM (1979) Diel vertical migration by
Salpa aspera and its potential for large-scale particulate organic matter transport to the
deep sea. Mar Biol 53:249-255
Wilson DW (1992) Interactions of ocean currents and diel migrators at a seamount in the central
north Pacific Ocean. PhD Dissertation, University of Hawaii, Honolulu, HI
Wootton RJ (1998) Feeding. In: Wootton RJ (ed) Ecology of Teleost Fishes. Kluwer Academic
Publishers, Dordrecht
Youngbluth M, Sørnes T, Hosia A, Stemmann L (2008) Vertical distribution and relative
abundance of gelatinous zooplankton, in situ observations near the Mid-Atlantic Ridge.
Deep-Sea Res (2 Top Stud Oceanogr) 55:119-125
77
Table 1. Zooplankton body type, corresponding geometric shapes, and body volume equations used for diet analysis of
Bathylagus euryops
Body type
Prolate spheroid
Cylinder
Shape
Taxa
Copepods,
Ostracods,
Ctenophores
Volume equation
π (l)(d2)
6
Body volume equation
π (prosome/body length µm)(prosome/body
width µm)2
6
Larvaceans,
Chaetognaths
π (d2)(h)
π (body width µm)2(total length µm)
4
4
Amphipods
π (d2)(h)
4
π (prosome width µm)2(total length µm)
4
78
Table 2. Gastric-evacuation rates and daily-ration estimates of Bathylagus euryops at 3⁰ C, 5⁰ C, 7⁰ C, and 9⁰ C
along the northern Mid-Atlantic Ridge in June 2004. Daily-ration estimates are mg wet weight of prey consumed
per day and percent of the average body wet weight of B. euryops consumed per day. Daily rations were not
generated for 7⁰ C and 9⁰ C in the Azorean Zone as specimens were not observed there.
Gastric Evacuation Rate
Temp
Low
High
3⁰ C
0.017
0.074
5⁰ C
0.034
0.093
7⁰ C
0.059
0.116
9⁰ C
0.099
0.145
Temp
3⁰ C
5⁰ C
7⁰ C
9⁰ C
Reykjanes Ridge
Daily Ration
Daily Ration
(mg per day)
(% weight per day)
Low
High
Low
High
4.99
22.31
0.01
0.05
10.28
27.91
0.02
0.07
17.62
34.92
0.04
0.08
29.73
62.47
0.07
0.15
Temp
3⁰ C
5⁰ C
7⁰ C
9⁰ C
Faraday Seamount Zone
Daily Ration
Daily Ration
(mg per day)
(% weight per day)
Low
High
Low
High
33.28
148.88
0.08
0.36
68.62
186.25
0.16
0.45
117.6
233.02
0.28
0.56
198.37
291.72
0.47
0.70
79
Temp
3⁰ C
5⁰ C
7⁰ C
9⁰ C
Charlie-Gibbs Fracture Zone
Daily Ration
Daily Ration
(mg per day)
(% weight per day)
Low
High
Low
High
11.95
53.43
0.03
0.13
24.63
66.85
0.06
0.16
42.21
83.63
0.10
0.20
71.2
104.7
0.17
0.25
Temp
3⁰ C
5⁰ C
7⁰ C
9⁰ C
Azorean Zone
Daily Ration
Daily Ration
(mg per day)
(% weight per day)
Low
High
Low
High
39.05
174.67
0.09
0.42
80.51
218.52
0.19
0.52
-
Figure 1. Trawl sampling stations for Leg 1 of 2004 R/V G.O. Sars MAR-ECO expedition to the MidAtlantic Ridge, from Iceland to the Azores (adapted from Sutton et al., 2008).
80
Figure 2. (A) Percent composition by wet weight in mg and (B) percent frequency of occurrence of prey
groups in the diet of Bathylagus euryops. Dietary indices derived from samples (n = 221) collected during the
2004 R/V G.O. Sars MAR-ECO research expedition to the northern Mid-Atlantic Ridge. Standard error
estimates, represented by error bars, were calculated from cluster sampling variance estimates.
81
Figure 3. Size frequency of Bathylagus euryops sampled along the northern Mid-Atlantic Ridge in
June, 2004 and used for dietary analysis.
82
A)
B)
Figure 4. (A) Scree plot depicting average distance between clusters versus the number of clusters which was
used to identify the number of clusters into which 30-mm (SL) size-classees of Bathylagus euryops should be
grouped (three size groups were selected since the curve leveled out at four or more clusters), (B) dendrogram
representing the relationships among the diet compositions of 30-mm (SL) size-classes of Bathylagus euryops.
Height represents the coefficient used as a measure of dissimilarity among size-classes.
83
Figure 5. Canonical correspondence analysis (CCA) biplot for Bathylagus euryops diet along the northern MidAtlantic Ridge, from Iceland to the Azores, in June of 2004. The canonical axes represent linear combinations of
three explanatory variables (fish size, ridge section, and depth zone), of which fish size and ridge section were
significant at α = 0.1. Fish size was represented by a small size class (35-95 mm SL), a medium size class (95-155
mm SL), and a large size class (>155 mmSL). Ridge section was represented by a northern ridge section
(Reykjanes Ridge and Charlie-Gibbs Fracture Zone) and a southern ridge section (Faraday Seamount Zone and
Azorean Zone). Depth zone was represented by the five depth strata sampled, including DZ1 (0-200 m), DZ2
(200-750 m), DZ3 (750-1500 m), DZ4 (1500-2300 m), and DZ5 (>2300 m).
84
A)
B)
C)
Figure 6. Diet composition (percent weight) of Bathylagus euryops collected along the northern Mid-Atlantic
Ridge presented by (A) small size class, (B) medium size class, and (C) large size class. n represents the total
number of specimens. Error bars represent standard error of the percent weight values of each of the prey types in
the small size class of Bathylagus euryops.
85
Figure 7. Diet composition (percent wet weight) of Bathylagus euryops collected along the northern Mid-Atlantic
Ridge relative to northern ridge sections (Reykjanes Ridge and Charlie-Gibbs Fracture Zone; top panel) and
southern ridge sections (Faraday Seamount Zone and Azorean Zone; bottom panel). Error bars represent standard
error of the percent weight values of each of the prey types relative to the northern and southern ridge sections.
86
THESIS CONCLUSIONS
87
Determining distribution and feeding patterns in abundant deep-sea fauna, such as
Bathylagus euryops, represents a critical component to establishing baseline data for which
future studies can compare to. The science of the deep sea has mostly operated under the
assumption of stable environments and unchanging situations over large areas (Herring, 2002).
However, we now know that the future scenario of climate change will influence all areas of the
marine realm. An increase in temperature over the next 100 years, predicted to be between 1.4
and 5.8 ⁰ C (IPCC, 2007), will have significant effects on the surface waters of the ocean, which
will then be transmitted to depth (Ramirez-Llodra et al., 2010). Furthermore, the expansions of
oceanic ‘dead zones’ will also have severe effects on the distribution and abundance of marine
life in the deep sea (Brewer and Peltzer, 2009).
The distributional trends observed in this study highlighted a significant influence of both
geographic location and depth on the abundance of B. euryops. The likelihood that climate
change will influence the distribution of B. euryops is high if current trends in elevated oceanic
warming and oxygen minimum zones persist. The ontogenetic vertical migration that was
postulated in the present study is one such mechanistic process that will be greatly affected by
the warming of the world’s oceans, and in turn, will influence the overall distribution of B.
euryops in the North Atlantic. As outlined in this study, it is likely that the ontogenetic
migration to the meso- and bathypelagic zones is preceded by the hatching of eggs in the
epipelagic zone. The continued warming of temperatures in the surface waters of the North
Atlantic may therefore expose the eggs and larvae of B. euryops to an expanded geographic
range of lethal temperatures. Considering the importance of depth on the distribution of B.
euryops, another important ramification of climate change is the expansion oxygen minimum
zones in the deep pelagic realm. The decline in oxygen in the deep pelagial is a result of lower
88
sea-surface oxygen concentrations, reduced ventilation of the meso- and bathypelagic zones from
ocean warming (Diaz et al., 2008; Stramma et al., 2008), and local eutrophication events, all of
which lead to an expansion of oxygen minimum zones (Brewer and Peltzer, 2009). The impact
of these oxygen minimum zones may impose a physiological strain on B. euryops in areas in the
water column where they are most abundant. These physiological strains would likely impair
performance of B. euryops and would require energy that would otherwise be used for
locomotion, predation, reproduction, or dealing with other environmental stressors (Brewer and
Peltzer, 2009).
Smith et al. (2008) reviewed the potential impact of climate change on deep-sea taxa and
concluded that organisms in the deep sea are highly vulnerable to temporal and spatial
perturbations in food availability and other environmental conditions. The deep sea relies mostly
on the downward flux of surface primary production and thus is often considered a food limited
environment. Changes in the quality and quantity of production in the epipelagic zone may
therefore have pronounced effects on deep-sea communities (Smith et al., 2008). For example,
one of the potential major consequences of climate change for B. euryops is a distributional and
abundance shift of arguably its most important food item, Calanus finmarchicus. Beaugrand et
al. (2003) predicted that in the North Atlantic, a shift in the distribution and abundance between
the northern C. finmarchicus and the southern Calanus helgolandicus can be expected. As
temperature warms, the boundary between the two shifts north and where their distributions
overlap, C. helgolandicus becomes comparatively more important (Drinkwater, 2005). Results
from this study have illustrated that the life cycle of C. finmarchicus appears to be linked with
trends in the feeding ecology of B. euryops. If a profound distributional and abundance
northward shift occurs for a major prey group of B. euryops, it is certainly plausible that climate
89
change will negatively affect the feeding ecology of B. euryops. While these predictions remain
highly uncertain, I believe they are a reasonable approximation of how climate change will
influence B. euryops throughout the mid-oceanic North Atlantic Ocean.
Overall, the consequence of the disturbance of deep-sea ecosystems is relatively
unknown and difficult to predict. However, the result of continued and widespread disturbances
would most certainly be severe and have long term effects on deep-sea fauna. The shear
abundance and biomass of Bathylagus euryops along the northern Mid-Atlantic Ridge, and the
subsequent distributional and feeding patterns observed in this study indicate that this species
plays an important role in the ecosystem of the mid-oceanic North Atlantic. Bathylagus euryops
is an important prey item for commercially fished species like orange roughy, Greenland halibut,
and alfonsino (Mauchline and Gordon, 1984; Chambers and Dick, 2007; Horn et al., 2010) and
the findings in this study can help in the management of these fishes that are highly vulnerable to
overfishing. Furthermore, the patterns observed in this study can be used to ultimately predict
future implications of environmental variability on the distribution and feeding ecology of B.
euryops, which can have cascading effects on both prey and predator populations.
90
REFERENCES
Brewer, P.G., Peltzer, E.T., 2009. Limits to marine life. Science 324:347-348.
Chambers, C.A., Dick, T.A., 2007. Using environmental variables to predict the structure of
deep-sea Arctic fish communities: Implications for food web construction. Arctic,
Antarctic, and Alpine Research 39:2-9.
Diaz, R.J., Rosenberg, R., 2008. Spreading dead zones and consequences for marine
ecosystems. Science 321:926-928.
Drinkwater, K.F., 2005. The response of Atlantic cod (Gadus morhua) to future climate change.
ICES Journal of Marine Science 62:1327-1337.
Herring, P., 2002. The deep-sea dimension. In: Crawley, M.J., Little, C., Southwood, T.R.E.,
Ulfstrand, S. (Eds.), The Biology of the Deep Ocean, Oxford University Press, New
York, pp. 1-26.
Horn, P.L., Forman, J., Dunn, M.R., 2010. Feeding habits of alfonsino Beryx splendens. Journal
Fish Biology 76:2382-2400
IPCC (International Panel on Climate Change), 2007. Climate change 2007: Mitigation.
Contribution of Working Group III to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge.
Mauchline, J., Gordon, J.D.M., 1984. Occurrence and feeding of berycomorphid and
percomorphid teleost fish in the Rockall Trough. Journal du Counseil – Conseil
International pour l’exploration de la mer, 41:239-247.
Ramirez-Llodra, E., Brandt, A., Danovaro, R., De Mol, B., Escobar, E., German, C.R., Levin,
L.A., Martinez Arbizu, P., Menot, L., Buhl-Mortensen, P., Narayanaswamy, B.E., Smith,
C.R., Tittensor, D.P., Tyler, P.A., Vanreusel, A., Vecchione, M., 2010. Deep, diverse,
and definitely different: unique attributes of the world’s largest ecosystem.
Biogeosciences 7:2851-2899.
Smith, C.R., De Leo, F.C., Bernandino, A.F., Sweetman, A.K., Martinez Arbrizu, P., 2008.
Abyssal food limitation, ecosystem structure and climate change. Trends in Ecology and
Evolution 23:518-528.
Stramma, L., Johnson, G.C., Sprintall, J., Mohrholz, V., 2008. Expanding oxygen-minimum
zones in the tropical oceans. Science 320:655-658.
91
VITA
Christopher J. Sweetman
Born in New Haven, Connecticut, 26 May 1984. Graduated from Daniel Hand High
School in Madison, Connecticut in 2002. Earned a Bachelor of Science in Marine and
Freshwater Biology from the University of New Hampshire in 2006. Worked as a molecular
genetics research specialist at the University of Connecticut from 2006-2008. Entered the
master’s program at the College of William & Mary in the School of Marine Science in 2008.
92